Transcripts of the monetary policymaking body of the Federal Reserve from 2002–2008.

  • Good afternoon, everybody. At the outset, I would like to welcome President Yellen and President-elect Lacker to their first FOMC meeting in their new roles. Janet Yellen, as all of you know—or at least those of us on this side of the table—served on the Board from August ’94 to February ’97. Jeff Lacker has been on the staff of the Federal Reserve Bank of Richmond since 1989 and has served as Director of Research since 1999. Welcome to the table.

    On a less happy note, Virgil Mattingly has decided that his apprenticeship here as the General Counsel of the Board and the FOMC for the last fifteen years is coming to an end. Virgil has been in the Board’s Legal Division for thirty years. Virgil, your absence will be duly noted—I don’t need to tell you nor do I have to tell any of the members around this table about the value of the contributions you have made and that have been felt implicitly in the deliberations of this organization. The fact that we hear from you so rarely at these meetings is testimony of how well you keep us out of trouble! [Laughter] And that, in my judgment, is the fundamental role of the General Counsel. We wish you well in whatever activities you choose to pursue, and we trust you will come back and visit with us on occasion just to say hello.

  • After one year! [Laughter] [Applause]

  • Virgil always gets the last legal zing. [Laughter] It has been a long run; you’ve been here as General Counsel of the Committee for almost my full term, and it’s getting long in the tooth also! Would somebody like to move approval of the minutes for the meeting of May 4?

  • Approved without objection. We will now move to the special presentation on prospective external adjustments, and I call on Karen Johnson.

  • Thank you, Mr. Chairman. You should have a packet of charts labeled “U.S. External Adjustment” in front of you. Our topic today is U.S. external adjustment. The “adjustment” that is our focus starts with the recognition that the United States is running a large and growing deficit vis-à-vis the rest of the world in the use of goods and services—a deficit that makes obvious our interdependency with foreign economic behavior. As a result, some analysts are concerned that a change in foreign behavior at some point may have unwelcome consequences for us. But in today’s interdependent global economy, such vulnerability is present even for those in external balance or surplus. In our presentation today, we will analyze the factors behind our growing external deficit and consider what circumstances might foster its reversal. Of particular concern will be the implications of various alternatives for U.S. asset prices, especially the foreign exchange value of the dollar.

    One direct measure of our external deficit (shown in the top panel) is the balance of payments concept termed the “current account balance”—that is, the sum of the trade balance, net investment income, and net foreign transfers. For many years the U.S. current account balance has been driven for the most part by the trade balance. Both balances began a period of substantial decline in 1996 that was only briefly interrupted by the recession in 2001.

    The price-adjusted exchange value of the dollar in terms of the currencies of a large number of our trading partners (the dotted blue line) is central to our story for two reasons. First, the dollar exchange rate is an important price variable in most, if not all, of the significant demand-and-supply relationships in our complex story. Second, the dollar exchange rate is part of the transmission mechanism of U.S. monetary policy and influences U.S. inflation determination. Note that the dollar began to appreciate from a low point in early 1995—in advance of the downturn in the trade balance—and generally continued to rise until early 2002. The nominal exchange value of the dollar in terms of the major foreign currencies (the green line in the middle left panel) has come even further off its 2002 peak than has the more inclusive broad real dollar, nearly returning to its 1995 level. This difference reflects the fact that, in terms of the currencies of our other important trading partners (the dashed red line), the dollar has risen somewhat in nominal terms since early 2002. The dollar exchange rates of the major currencies have tended to be more flexible and more market driven than some of the OITP currencies, such as the Chinese renminbi; and the dollar has appreciated against some of these other currencies, in particular the Mexican peso.

    Persistent current account deficits cumulate to the net international investment position, shown to the right, which measures our growing stock of indebtedness to the rest of the world. This large and growing stock of claims on the United States must be held at all times, and one intuitive perspective on the sustainability of our external position is that this stock cannot increase relative to our GDP forever. Every import and export transaction in the current account has a financial counterpart. Except for measurement error, the current account balance equals the net of all the international financial transactions. There are, of course, huge volumes of cross-border transactions between assets—unrelated to trade finance and driven by complex incentives of risk and return—that have consequences for asset prices, including exchange rates. Recent data for elements of those financial flows, reported in the bottom left panel, show that private foreign inflows to purchase U.S. securities (line 3) are very large, particularly in relation to private U.S. outflows to purchase foreign securities (line 4) and net direct investment (line 5). Moreover, these flows increased sharply in the first quarter. Nevertheless, foreign official inflows (line 2), which had about doubled in 2003, rose significantly further earlier this year.

    From a macroeconomic perspective and with the public and private sectors combined, the extent to which the United States invests more than it saves corresponds to resources borrowed from abroad—the blue shaded region labeled “net foreign lending” in the bottom right panel. This is the NIPA analogue to the current account balance. As the rise in U.S. net domestic investment in the 1990s outpaced that in net saving, our external deficit rose. As our saving subsequently declined sharply, the external deficit widened further.

    Unless and until the trade balance begins to narrow, adjustment is not yet occurring. Indeed, as was argued in the background paper circulated to the Committee, stability in the ratio of our foreign indebtedness to nominal GDP requires that the trade deficit eventually return to near zero. U.S. trade data are presented at the top of your next exhibit. The yellow shaded area represents the trade deficit, which has reached a record level. To the right of the vertical line, which is positioned at the latest actual data point, we have shown a simple extrapolation of imports and exports—based on standard partial equilibrium trade equations—for the case in which real GDP here and abroad grows at potential and the real value of the dollar remains at its current level. With imports already substantially larger than exports and with foreign potential growth reasonably robust but slightly less than U.S. potential growth, the change in imports tends to exceed that in exports, and so in this case there is no adjustment through 2010. The green bars measure the change in the trade balance. For the trade balance to narrow significantly, exports would need to grow rapidly relative to import growth for an extended period.

    The middle panel provides a summary of the composition and magnitude of cross- border financial flows for the same period, expressed as a percent of GDP. The trade balance above accounts for most of the change in the current account balance, shown in yellow. U.S. private financial outflows, in tan, also must be financed; and the sum of these two is the counterpart to foreign private financial inflows (in blue) plus net official flows (in red). The extrapolation through 2010 incorporates the trade outcome from the top panel and plausible staff projections of U.S. and foreign private financial flows. Of course, private flows could behave differently, depending on the incentives perceived by investors here and abroad. To the extent that the flows shown in blue do not fully finance the current account deficit plus U.S. private outflows, foreign official intervention may fill the remaining gap, as shown here. Otherwise, exchange rates, other prices, and incomes here and abroad would change, resulting in a current account position and private flows that would satisfy the balance of payments norms.

    The bottom two panels provide some detail on recent stocks of dollar holdings by foreign officials (the left panel) and foreign private investors (the right panel) that correspond to the flows in the middle panel. Estimates based on data through April show that, of the total foreign official holdings of more than $1,600 billion (line 1), about two-thirds consist of Treasury securities (line 2). The end of 2001 was very near the time of the most recent peak in the value of the dollar. Changes from then through April are shown in the column to the right and reveal that the change in the total is nearly accounted for by the change in holdings of “selected” Asia (line 3)—economies whose monetary authorities have engaged in substantial foreign exchange intervention. The incentives behind dollar acquisition by foreign officials, for the most part, relate to their objectives for their respective exchange rates and the competitiveness of their exports.

    Foreign private holdings are significantly in fixed-income securities—lines 1, 2, and 3 in the right panel. These investors are thus exposed to dollar exchange rate risk and to U.S. market interest rate risk. Foreign holdings of U.S. equity (line 4) are also large but entail somewhat different risks. Over the period since the dollar began its appreciation in 1995, foreign investors have more than doubled their holdings of each of these categories of assets. Nevertheless, because U.S. securities markets are very large and quite liquid, changes in preferences on the part of foreign investors among these categories of assets need not have major implications for their prices. However, as long as the U.S. current account remains in deficit, net foreign claims on U.S. assets must continue to rise in total. Any diminution in the appetite of investors for U.S. assets would first trigger valuation changes, primarily via the exchange rate, that would lower the foreign-currency denominated value of their total claims on the United States.

    Such a change in investor appetite could well be the first step in the external adjustment process. Some characteristics of an “orderly” adjustment are listed in the top left panel of your next exhibit. Although there are a great many possible alternative paths for adjustment that could be termed “orderly,” in the sense that financial markets would continue to function within normal ranges, such an outcome seems more likely to occur if investors are attracted by improved returns abroad rather than discouraged by unfavorable changes in U.S. prospects. Various specifics along these paths would differ, but all would be characterized by continued, albeit declining, net financial inflows into the U.S. economy. In addition, because lasting correction of the U.S. external deficit cannot be achieved through cyclical slowing of U.S. GDP growth to below potential and the associated temporary restraint on imports, external adjustment such that exports increase more than imports almost certainly requires dollar depreciation.

    The panel to the right expresses the data on foreign holdings of U.S. assets from the previous exhibit as shares relative to the total outstanding amount for each category. Foreign holdings of U.S. Treasury securities (line 1) amount to nearly 50 percent of the outstanding stock, with the official share about 30 percent and the private share (not shown) about 20 percent. Foreign holdings are more than 10 percent of agency securities. The percent held abroad is only slightly higher for equities, with 12 percent of U.S. equities owned by foreign investors, virtually all of whom are private. These numbers are large, and they have risen over time. But moderate financial inflows could continue over time even if these shares stabilize or edge down.

    There appears to be ample scope for continued financial inflows, as reflected in the bottom left panel. For most of the foreign countries shown, in December 2002 domestic securities made up most of their total bond and equity holdings, in many cases more than 80 percent. Moreover, U.S. securities are generally just a moderate portion of their nondomestic holdings. Accordingly, it does not seem to be the case that foreign portfolios are already overly invested in U.S. assets.

    If stronger demand for U.S. exports is to contribute to adjustment, from which countries might it come? The panel on the right reports that the economies with the largest average share of U.S. exports in 2003 were Canada and western Europe (lines 1 and 2) followed by Mexico (line 3). These regions accounted for a slightly smaller share of the increase in U.S. exports over the past two years, shown in the second column. Note the jump in the share for China and Hong Kong (line 7) when viewed from the perspective of the change in exports. This increase is somewhat offset by the drop in Japan’s share (line 5). Taken together, our Asian trading partners (lines 4, 5, 7, and 8) accounted for 30 percent of the gain in exports. Income growth in Asia is likely to remain quite vigorous, with domestic demand a major source of strength. To a lesser degree, this outcome is likely for Canada and Mexico. Strong domestic demand in those regions will be necessary if U.S. exports are to accelerate.

    The panel at the top left of your next exhibit is constructed to illustrate the sense in which there is a tradeoff between relative growth here and abroad and the minimum pace of real dollar depreciation required for the trade deficit to narrow. The line separating the shaded and nonshaded regions was calculated using standard partial equilibrium trade equations, and its position depends importantly on the current starting conditions, in particular the initial size of imports and exports. Given those quantities today, for any foreign minus U.S. growth gap (the horizontal axis) the unshaded area indicates what the annual rate of dollar depreciation (the vertical axis) would have to be over the near term for the U.S. trade deficit to begin to adjust, following the usual lags. In 2003, we were well above the line, with a growth gap of minus 1½ percentage points and real dollar depreciation of 8½ percent. The panel shows that the stronger relative foreign growth is, the less dollar depreciation would be required. However, even for growth outcomes very favorable to foreign activity, some dollar depreciation would be needed to get adjustment started. The staff estimates that the gap for relative potential growth is currently about minus ½ percent, so we have plotted the corresponding point on the line, marked by the box, as indicative of how far to the right on the horizontal axis we might tend to be. At that growth gap, dollar depreciation of more than 8 percent per year would be needed to start the process of external adjustment.

    Overall, the implications for the U.S. economy of extended, gradual external adjustment—listed to the right—include that demand on the part of U.S. residents and foreigners shifts in favor of U.S.-made goods and services, thus boosting exports and reducing imports. In order for the United States no longer to depend so much on foreign saving, the extent to which U.S. domestic demand exceeds U.S. production potential must decline. Contractionary fiscal policy could contribute to restraining domestic demand and increasing net saving. Monetary policy would need to maintain full utilization of U.S. resources and may need to offset the expansionary implications of increased demand for U.S. goods. Significant resource shifts across production sectors may be needed, which could entail adjustment costs for the economy.

    We cannot rule out that adjustment will be disorderly. In that case the primary concerns are about conditions in asset markets, where there could be abrupt price changes, increased volatility, and impaired liquidity, triggered by changed sentiment on the part of foreign and domestic investors about the attractiveness of U.S. assets. Because of the role of the exchange rate in cross-border transactions, dollar depreciation is likely to be at the center of any disorderly event. Given that some depreciation is part of virtually any external adjustment, depreciation in the case of a disorderly outcome would likely have to be large in magnitude and abrupt in pace. With global asset markets now highly linked, any abnormal developments in U.S. markets would be likely to spill over to asset markets elsewhere.

    Where are the effects of such asset-price swings likely to matter most? The bottom left panel contains data on balance sheet risk faced by U.S. corporations via their debt denominated in foreign currency. Foreign currency issuance is just 4 percent of the total outstanding, and a substantial portion of that foreign currency risk may well have been hedged in the derivatives market. So this channel is not likely to have significant consequences. Foreign holdings of claims on the United States are shown by region in the panel to the right. In European portfolios, especially those in the United Kingdom, and in Canadian portfolios, holdings of U.S. assets are quite large relative to their respective GDPs. Negative balance sheet effects through declines in the exchange rate or U.S. asset prices could be expected to significantly impair economic activity in those regions and to be not inconsequential for the other countries shown. Joe Gagnon will continue our presentation.

  • Your next exhibit considers the implications for both the U.S. and foreign economies of an abrupt depreciation of the dollar that could be associated with a disorderly adjustment. As described in the top left panel, we examine three scenarios that build on each other using the staff’s FRB/Global model. We note at the outset that these scenarios are illustrative of some of the issues that may arise and should not be viewed as forecasts or even as particularly likely outcomes.

    In scenario 1, there is an exogenous permanent drop in demand for dollar assets that would cause the broad real index of the dollar’s value to decline 30 percent over the next two quarters in the absence of any responses in domestic and foreign interest rates. The only decline of this magnitude in the history of the broad real index occurred over a three-year period from 1985 to 1988. In this and subsequent scenarios, we assume that U.S. and foreign policy interest rates follow a Taylor rule. The dotted black line in the panel to the right shows that the dollar actually depreciates less than the full 30 percent in this scenario, as higher U.S. and lower foreign interest rates offset some of the depreciation shock. Column 1 of the middle panel shows that U.S. GDP (line 1) is about 2 percent higher after six quarters. The effect on the level of U.S. output peaks at just over 2 percent in early 2006 and then gradually unwinds. Higher output is more than accounted for by net exports (line 3), as domestic demand (line 2) declines. By the end of 2006, core PCE prices (line 4) are 1½ percent higher. This price increase is stretched out over ten quarters, so the annualized inflation rate never rises more than a percentage point above baseline. Most of this rise reflects the direct pass-through of higher import prices (line 5), but a little is attributable to the rise of output relative to potential. On its own, this shock would not be considered disorderly, and it reflects a relatively benign response to such an abrupt dollar depreciation. Nevertheless, the negative effect on trade-weighted foreign GDP (the dotted black line in the bottom left panel) is a bit larger than the positive effect on U.S. GDP as policy rates in many important U.S. trading partners hit the zero bound on nominal interest rates. Foreign real consumption does not fall as much as foreign output because dollar depreciation lowers the cost of imports to foreign consumers. The opposite effect is at work in the United States, showing that even though dollar depreciation is good for U.S. production, it does have adverse effects on U.S. consumers.

    In scenario 2, we posit that this abrupt dollar depreciation is associated with a widespread loss of confidence in U.S. economic prospects that leads to a 250 basis point increase in the risk premium on equities and long-term bonds in the United States. The higher premium and lower expected profits cause U.S. equity prices to fall about 50 percent. Because of the importance of the U.S. economy and U.S. financial markets in the global economy, it is likely that such a large drop in U.S. equity prices would spread to foreign markets. The overall effect on foreign financial markets would likely be muted by safe-haven flows to government bonds in major foreign countries. Thus, we set the size of the foreign financial shock to have half as much effect on foreign output as the U.S. financial shock has on U.S. output. Asset-price declines add contractionary impulses to the simulation, and now U.S. GDP, shown in columns 3 and 4 of the middle panel, declines on balance. At the same time, the upward pressure on U.S. prices (line 4) is reduced a bit, owing to slack in resource utilization. As shown by the dashed red line in the bottom left panel, asset-price declines in foreign economies and spillovers from the United States have a substantial contractionary effect on our trading partners. Policy rates in most of these countries drop to zero early in the simulation and remain there.

    Scenario 3 considers the possibility that, when policy rates hit zero, foreign central banks might engage in quantitative easing or other nontraditional policy actions to stimulate their economies. Because the channels for such policies are not built into the model, we chose to relax the zero bound on policy interest rates and allow these rates to become negative, still following a Taylor rule. As shown by the solid green line in the top right panel, easier monetary policy abroad leads to less dollar depreciation. Nevertheless, the level of U.S. GDP (line 1, columns 5 and 6 of the middle panel) is essentially identical to that in scenario 2. The reduced stimulus to U.S. GDP from a smaller depreciation is roughly offset by the reduced drag from a smaller contraction in foreign GDP, shown by the solid green line in the bottom left panel. The smaller depreciation helps to damp the rise in U.S. prices (line 4 of the middle panel). The bottom right panel displays the classic J-curve effect of a depreciation on the trade balance. Most of the effects are completed by early 2006. The trade deficit narrows about 1½ percent of GDP from the exchange rate shock alone and about 2 percent of GDP when other asset prices also drop. These scenarios engender a substantial degree of external adjustment, though not enough to eliminate the trade deficit, which is currently around 5 percent of GDP.

    All together, these simulations clearly indicate that disorderly adjustment of the U.S. external balances is likely to have a very contractionary effect on foreign economic activity. A large dollar depreciation—which is likely to be at the heart of any disorderly adjustment—provides a substantial stimulus to U.S. exports that offsets much of the negative effects of any financial distress on U.S. activity. Meanwhile, lower U.S. imports are a direct drag on foreign economic activity.

    In light of the dollar depreciation that occurred over the past two years, exhibit 6 explores whether external adjustment is already under way. The last major period of external adjustment in the United States occurred during the late 1980s. We find it instructive to compare the recent experience with that of the 1980s. The dashed line in the top left panel displays the rise and fall of the broad real dollar between 1979 and 1990. The solid red line displays the broad real dollar in recent years, with the peak of February 2002 aligned with the earlier peak of February 1985. The amplitude of the cycle was greater during the 1980s, but the broad contours look similar. The recent interruption in the depreciation is more pronounced than in any comparable period in 1985-87. Should the recent break in the dollar’s downward trend persist, it would be reasonable to expect any adjustment to be interrupted. The top right panel displays the trade balance over the earlier and later periods. Despite the smaller swing in the dollar in the current episode, the trade deficit as a percent of GDP has widened more than during the 1980s. Abstracting from the effects of the 2001 recession and 2002 recovery, there does appear to have been a flattening out of the trade deficit recently. In the 1980s, the trade deficit did not begin to narrow until more than two years after the dollar peaked—we are just now approaching the comparable point in the current episode. The middle panel shows that external adjustment in the late 1980s was associated with a sustained higher growth rate of real exports (the dashed blue line) relative to growth of real imports (the solid green line). In the current episode, the growth rate of real imports has exceeded that of real exports continuously from 1997 through late last year, but the gap has now closed. Adjustment will require a further rise in the export growth rate relative to that of imports.

    The bottom left panel displays weighted foreign real GDP growth minus U.S. GDP growth. As shown by the dashed line, foreign economies grew faster than the U.S. economy in the late 1980s. Robust foreign growth undoubtedly contributed to a relatively smooth adjustment process during this period. In contrast, during recent quarters, foreign economies on average have grown more slowly than the United States (the solid red line), and we are not projecting a reversal of this relationship in the near future. The bottom right panel gives some indication that the willingness of international investors to fund our widening external deficits may be diminishing. Net private inflows of securities have failed to continue rising since the dollar’s peak in early 2002; foreign official inflows have taken up the slack to finance the continued growth of the current account deficit. A similar rise in official inflows occurred at the beginning of the adjustment process in the 1980s.

    The top left panel of your final exhibit compares the recent behavior of import prices with the previous episode of adjustment. As shown by the red line, import prices have turned up with the dollar depreciation and the commodity price increases of the past couple of years. But the rate of increase is much smaller than that observed in the late 1980s. Moreover, unless the dollar depreciates significantly further, we expect that import-price inflation will drift down over the next year or so. The panel to the right shows that there was not a substantial increase in core consumer price inflation associated with the 1980s adjustment. We do not expect a major increase in core PCE inflation in the near future. Overall, the evidence to date points to a pause in the widening of the trade deficit, but that is far from a significant adjustment. Indeed, as Karen mentioned previously, if the dollar remains near its recent level we would project a renewed widening of the trade deficit.

    In conclusion, we believe that recent levels of the U.S. external deficits are not sustainable indefinitely. However, we know little about the path that adjustment will ultimately follow, including how long our large trade deficits can be financed. The depreciation of the dollar in 2002 and 2003 has helped to slow the widening of the trade deficit, but there is no evidence that a sustained or significant adjustment has begun yet. Assuming that U.S. and foreign output remain close to our estimates of potential, a substantial further dollar depreciation is required just to get adjustment started. The relatively orderly adjustment of the late 1980s was associated with an acceleration of foreign economic activity and brighter investment prospects abroad. On the other hand, a disorderly adjustment process would more likely be associated with a loss of confidence in U.S. economic policies and prospects. In a disorderly adjustment, the contractionary effects on output could well be greater for foreign economies than for the U.S. economy. While asset-price declines tend to depress output both at home and abroad, dollar depreciation tends to boost U.S. production and damp foreign production. The effect on U.S. inflation of even a large depreciation is likely to be quite modest. Linda Goldberg will now continue our presentation.

  • I will be referring to the separate package of exhibits that you should also have in front of you. My portion of this briefing on U.S. external adjustment focuses on the exchange rate and trade exposure of U.S. industries. Some scenarios for closing the U.S. trade deficit involve depreciations of the dollar exchange rate against other currencies. Yet such depreciations do not affect all U.S. producers similarly. My background material provides various industry-level details. In my remarks today, I highlight four broad points. First, the most trade-oriented industries of the United States are in the manufacturing sector and account for almost half of manufacturing employment. There are also some high-trade-oriented raw-materials industries. Second, these trade-oriented industries are the ones expected to be stimulated most by dollar depreciation. Moreover, the currency that the dollar depreciates against is important; U.S. industries experience greater stimulus when the dollar depreciates against the euro or the yen, for example, than against the yuan. Third, for many industries, rates of exchange rate pass-through into import prices have been relatively stable over past decades. While there is evidence of reduced import-price responsiveness to exchange rates in some commodities, such declines in pass-through may have been temporary. My fourth point is that a few high-trade-oriented industries in the United States may account for the bulk of the import and export adjustments induced by dollar depreciation.

    We begin by turning to the levels of international trade exposure of specific U.S. industries. We discuss three forms of trade exposure: (1) producer export orientation— the size of exports as a share of producer shipments; (2) the extent to which foreign producers have penetrated U.S. markets; and (3) industry use of imported inputs in the production processes—the share of production costs attributable to imported components and machines. We define high-trade-oriented industries as having export orientation above 20 percent and import penetration above 20 percent. These industries, shown in exhibit 1, are concentrated in our manufacturing sector and include chemicals; machinery excluding electrical; computers and electronics; electrical equipment; transportation equipment; miscellaneous manufacturing, such as toys and jewelry; and leather products, which is a small industry in the United States. Taken together, these industries account for 44 percent of U.S. manufacturing jobs. Outside of manufacturing, some raw materials sectors are also heavily trade oriented either in exports or imports. Moreover, if we broaden our criteria to allow a more narrow focus on import penetration, we can also include apparel, primary metal manufacturing, and furniture and fixtures. The dark blue bars in exhibit 1 show industry export shares, and the red bars display import penetration of these high-trade-oriented industries; both bars suggest producer revenue exposure to international trade. The yellow bars show the role of imported components in industry costs. The use of imported inputs means that producer costs rise when the dollar depreciates, as long as there is some pass-through of exchange rate changes into the prices of these imported components. To the extent that the same producers are exposed to the same currencies on the revenue and cost sides of their balance sheets, producer profits are partially hedged against currency fluctuations. All manufacturing and nonmanufacturing industries have revenue exposure in excess of their cost exposure to international trade. This leads to the expectation that a trade- weighted dollar depreciation would be stimulative, on average, for U.S. industries.

    Exhibit 2 presents the countries that are the destinations for aggregate U.S. exports, as well as those that are the sources of our aggregate imports of goods. Canada and Mexico together account for more than 30 percent of total U.S. exports and imports. The euro area accounts for 15 percent. Our trade imbalance with China is suggested by the fact that that country accounts for only 4 percent of U.S. exports but 13 percent of our imports. These shares of countries in overall U.S. trade transactions are similar to those used by the Federal Reserve Board in constructing the aggregate exchange rate indexes for the United States. However, not all industries are exposed to the same trade partner countries or, consequently, to the same bilateral exchange rate movements. Aggregate exchange rate indexes, therefore, will not reflect the changing value of the dollar from the perspective of specific U.S. industries.

    This point is made clear in exhibit 3, which shows the shares of the euro area, Japan, and China in the exports and imports of each of the most trade-oriented U.S. industries. As shown in the first column of the exhibit, the euro area is a key destination market for our exports. The implication is that the euro–dollar exchange rate figures prominently in the export competitiveness of, for example, chemical products, transportation equipment, and miscellaneous manufacturing products. In contrast, China is a relatively small destination market for U.S. manufactured goods but is larger in some import categories. The yuan–dollar exchange rate is significant mainly to the extent that it changes the competitiveness of U.S. producers relative to Chinese producers of goods being sold in the United States and abroad.

    A policy-relevant issue here is whether a dollar depreciation against the yuan would stimulate U.S. industries. A stimulus would most likely occur in those industries where Chinese producers compete head to head with U.S. producers. I address this point in exhibit 4, which presents the amount of U.S. manufacturing employment in industries with different degrees of import penetration by China. The top row of exhibit 4 reveals that about 70 percent of U.S. manufacturing jobs are in industries where “made in China” goods account for a low share (less than 5 percent) of overall goods consumed by U.S. households. The bottom row shows that high levels of Chinese import penetration are strongest in those U.S. industries that account for very few U.S. jobs (only 2 percent of manufacturing). There are few industries where “made in China” accounts for more than 20 percent of our consumption of particular goods. In the middle rows of exhibit 4 are the industries with both import penetration by Chinese producers and sizable U.S. manufacturing employment. Industries such as furniture, electronic components, photographic equipment, and computers and peripherals are those that can potentially experience a switch in demand from Chinese goods to U.S. goods and are therefore the ones most likely to gain from a dollar depreciation against the yuan.

    In evaluating industry exposure to dollar real exchange rate moves, let’s first consider the progression of the dollar against five currencies since 2000. Exhibit 5, on the next page, shows that the dollar had strengthened against the yen, the euro, and the Canadian dollar through early 2002, before beginning its descent. Currently, the dollar is close to its weakest level in four and a half years against the Canadian dollar and the euro; it has had a more limited descent against the yen. By contrast, the dollar strengthened overall against the peso and the yuan. We use these different paths of exchange rates together with the weights of our trade partner countries in each industry to derive industry-specific export-weighted or import-weighted exchange rates. Exhibit 6 shows the dollar depreciation experienced by high-trade-oriented U.S. industries since February 2002, separately presented from the perspectives of U.S. exporters and U.S. import-competing producers. The recent dollar depreciation for high-trade-oriented industries has ranged from 7 percent to a little over 18 percent for U.S. exporters, as shown in the first data column of exhibit 6. The corresponding depreciation for U.S. import-competing producers has ranged from about 4 percent to nearly 27 percent. A key point here is that, compared with other high-trade-oriented industries, the industries that have China and Korea and even Japan as large trading partners have experienced relatively less depreciation and smaller changes in competitive conditions since February 2002.

    As a final theme, in exhibit 7, I focus on exchange rate pass-through into U.S. import prices. This theme is central to knowing whether exchange rate changes can induce enough demand-switching out of imports and exports to make a dent in our trade balance. Pass-through is also important for understanding the potentially major distributional consequences induced by exchange rates within the United States. Exhibit 7 provides estimated rates of the pass-through of exchange rate changes into import prices for different bundles of imported commodities. The estimates in the left column are for regressions run using data spanning the late 1970s–early 1980s through the second quarter of 2004. The right-hand column estimates are for a more recent period, starting in 1990. The elasticities of import prices with respect to exchange rates that are statistically significant are in bold print. The first two rows of exhibit 7 show exchange rate pass-through into the prices of an aggregate bundle of United States imports. The historical relationship indicates that foreign producers generally absorb a high share of exchange rate changes in their profit margins in the short run (about 70 percent) and about half of exchange rate changes over the longer run (one year). A rule of thumb is that a 10 percent dollar depreciation raises U.S. dollar import prices by 5 percent. The stability of this relationship between exchange rate movements and import prices in the United States is actively debated, especially with regard to whether exchange rate pass-through has declined recently. The comparison between the full data sample estimates in the left column and the recent sample estimates in the right column is at the heart of this debate.

    Exhibit 7 highlights three important points. (1) The U.S. dollar prices of beverage, tobacco, and mineral fuel imports are relatively insensitive to changes in exchange rates, with food prices only marginally sensitive. (2) Exchange rate pass-through rates into the import prices of food, chemical products, and some manufactured goods have been stable over time. A 10 percent dollar depreciation increases import prices of these goods by a cumulative 4 to 7 percent within a year. (3) Pass-through may have declined on U.S. imports of machinery and transportation equipment; the relationship estimated using historical data consistently overpredicted import prices for this commodity in 2002 through mid-2003. This category of machinery and transportation equipment has a heavy weight in the aggregate bundle of U.S. imports, which explains why we estimate lower pass-through since the 1990s on the overall bundle of U.S. imports excluding fuels. Overall, in my judgment, it is premature to assume systematic and persistent declines in exchange rate pass-through into import prices. The recent decline in pass- through may have been temporary. Import-price observations from late 2003 into early 2004 are back within the bounds of the longer-run historical relationships. We may see foreign producers reaching a limit on the extent to which they let their profit margins absorb the adverse exchange rate movements that have occurred. Therefore, more time needs to pass before we conclude that exchange rate pass-through into import prices is lower than what is implied by the rule-of-thumb estimate.

    I conclude by combining these themes in a way that is directly relevant to the external adjustment process for the aggregate economy. Exhibit 8 shows the substantial role that the high-trade-oriented industries play in overall U.S. trade—accounting for more than 60 percent of our total exports and imports. These are the industries on the rightmost side and bottom of each pie chart. The significant pass-through of exchange rate changes into the import prices of these industries provides the scope for expenditure switching between domestic goods and imports. Such expenditure switching within the United States is particularly feasible in those industries where we continue to have a production and employment presence. In my opinion, these industries are likely to be the ones that bear the brunt of trade adjustment in the scenarios for closing U.S. external imbalances. Moreover, since industrialized countries are the dominant markets for our exports, adjustments to our international trade balance—and to U.S. jobs—will be most pronounced for changes in the dollar exchange rate against the euro and the yen rather than, for example, the yuan. Since Chinese imports compete directly with a much smaller portion of our products, dollar depreciation against the yuan would be less effective in the pursuit of a sizable shrinkage of our trade imbalance. Continual gradual declines in the dollar against the euro and the yen would lead to fewer U.S. imports and more U.S. exports. Changes in our trade balance will appear to be driven by the export adjustments. After cumulative large declines in the dollar, some U.S. producers might expand their market orientation to also focus on exporting instead of just U.S. markets, and this would then speed the overall adjustment of the trade balance. Although import quantities are changing in the background, too, the amount of total expenditure on imports will look relatively unchanged. Thank you for your attention and for the opportunity to present these points.

  • The last point that you made is particularly apt in the sense that we have a general tendency to evaluate these types of major adjustments in the context of historical relationships, and we don’t think in innovative terms largely because innovations are very difficult to anticipate. Let me just say that overall these presentations have provided an extraordinarily incisive view of the nature of the adjustment processes and a few of the potential avenues through which an ultimate solution can occur.

    In your pass-through analysis, you made no reference to hedging characteristics. We do know, for example, that in very recent years the outstanding volume of dollar–euro foreign exchange swaps has gone up very dramatically. Clearly, to the extent that that is taking place, the greater return on exports is a combined function of hedging plus actual pass-throughs. And even though we may see a very sharp contraction in implicit profit margins of exporters out of Europe to the United States, for example, or anywhere, from the point of view of the exporters those losses may be fully offset through the exchange market. And the exporters’ behavior doesn’t change until ultimately as time goes on they cannot avoid the depreciation of exchange rates. So this is a process that slows the extent of adjustment but cannot alter it. I understand the data are really quite inadequate for any industry analysis, but we do have aggregate data, and I was wondering if you have come to any conclusion as to what we can learn from that.

  • Let me start by mentioning the response of profits to exchange rate changes. First, at the industry level, if we use an aggregate trade-weighted exchange rate to look for the effects of exchange rates on profits of U.S. industries, we don’t see anything that is statistically significant. If we use industry-specific weights, then we do get a significant relationship between exchange rate changes and the profitability of U.S. firms.

    There is another effect of the dollar depreciation or appreciation on U.S. corporate profits that I didn’t mention—in addition to the hedging story, which comes through multinational activities—and that is that, when the dollar moves, there is pretty much a one-to-one translation effect that shows up on the profits of multinational activities. That means, for example, that when the dollar depreciates, the foreign currency profits of these firms rise in value relative to their U.S. dollar profits. And that shows up as at least a temporary blip in the overall profitability of these firms.

    I don’t have any direct evidence on hedging activity and the effect that has had on pass- through and profitability, but some of the implications of exchange rates for profits are reduced by the increasing role of imported components used in production, which one can think of as another way that producers hedge foreign exchange rates.

  • We do see that the goods imported into the United States in dollars from Western Europe show a far smaller inflation rate than would be implied by the one-to-one corresponding foreign exchange rate. That implies either a significant decline in profit margins for European exporters or a very substantial shorting in effect of the euro in relation to the dollar. I gather, as I mentioned before, that this is a very short term phenomenon. I presume you have decided that its impact, while it may show up for a year or two, is really not a relevant consideration, largely because the foreign exchange currency swaps almost never go beyond a couple of years.

  • Mr. Chairman, may I?

  • We have looked, partly at your suggestion, into whether the lags might have changed, because I think the mechanism you are talking about would likely show up as a longer pass-through lag. Obviously we don’t have enough data to determine if this is a recent change in the past couple of years; we can’t tell. We looked at whether the ’90s were different from the ’80s and couldn’t find any evidence of a longer lag time. The lags are pretty short.

  • Thank you. Looking at the broader question of the very important impact in terms of contractionary effects on foreign economies, are the calculations here made interactive? In other words, are we getting two static estimates, or are we getting a dynamic estimate—a feedback from the United States to foreign economies and back?

  • There would be dynamic feedbacks. We use the FRB/Global model, which will simultaneously capture both effects—from the United States on foreign economies and from foreign economies on this country. So, yes, it’s a property of the model that things that happen in the United States have a bigger effect on the rest of the world than vice-versa, in large part because in this exhibit we weighted the rest of the world by U.S. trade weights to give greater weight to countries that trade with us a lot, such as Canada and Mexico and Asian countries. If we had done the weighting by market or PPP-GDP, say, we would get a smaller effect.

  • All of this says, at the end of the day, that the U.S. current account deficit is rising as a percent of our GDP, or in a certain sense as a percent of world GDP, implying that we must be getting comparably increasing surpluses elsewhere in the world. Or put another way, this implies that the dispersion of the world’s current account balances is increasing. That is almost arithmetically necessary in the data that you have here. Are we seeing any signs of a slowdown in the globalization process? For example, in whatever time frame we wish to look, what do we see in the ratio of aggregate world exports to world GDP, aggregate external gross assets and liabilities as a percent of GDP, or any of these broader macro relationships that relate to the dispersion of current account balances worldwide? In short, granted the long delays on a lot of these types of data, which almost all come out of the IMF data system one way or another, is there any evidence that the rate of change here is slowing? What has been happening in the last couple of years, for example, to the data that are perhaps the most current—world exports as a percent of world GDP? How would you describe the quarterly pattern over the last several years?

  • To be honest, I don’t have those numbers in front of me.

  • We could certainly calculate world exports to world GDP. My guess is that it won’t have slowed.

  • It slowed down very dramatically a couple of years ago.

  • Well, the recession.

  • Yes, but has it come all the way back?

  • I can’t answer that exactly. Part of the problem, of course, is finding the units in which to add up world exports and world GDP; that requires us to use data sources that are not available on a timely basis. I don’t have world GDP for Q1; that number doesn’t exist.

  • You don’t? What a shock! [Laughter]

  • But what is true is that the global recession and recovery tended to be more intensive in goods than in services. They tended to be investment driven, at least in the United States and to some extent elsewhere. And trade is more goods intensive than is GDP, broadly speaking. So we saw imports and exports react more than in proportion to our GDP or to our weighted average of foreign GDP; and presumably the rest of the world, added up, experienced the same phenomenon. I would assume that we are still on something of an uptrend and are washing through a cyclical phase, but we can look at that and get back to you on it.

    Of the variables that relate to globalization in a somewhat more indirect way, the only ones that I know of that really slowed were the turnover measures on the foreign exchange market. I think, in part, the effect of the euro on foreign exchange turnover, as we measure it, probably explains why we saw declines in the volume of foreign exchange activity.

  • But isn’t that also a function of the rate of change in the exchange rate itself?

  • There were at least three components. One, there was an exchange rate effect. Two, there was the euro. And there was a third factor that had to do with the increased automation in terms of how the exchange market operated. Electronic brokering, for example, became very, very important.

  • It grosses up everything.

  • Exactly. The numbers were down about 20 percent from ’98 to ’01, but all of the signals that we are getting now suggest that the survey just concluded in ’04 will show another increase.

  • But the grossing doesn’t tell us what the consolidated change in globalization is. The trade balances do, obviously, and the balance sheet changes do because they are all directly related algebraically to the current account balances and, hence, to the degree of dispersion. There is a chart in exhibit 2 of Karen’s material, which has an extrapolation of the current account balances out to 2010, when it goes to over 9 percent of U.S. GDP. I’m looking at that and I’m saying, okay, why not? There’s nothing in this presentation that says “no, that can’t happen.” It merely says that someday, somewhere, somehow, that trend is going to change. Let me just ask a quick side question. How did you conclude in that particular chart, the central one in exhibit 2, that foreign private financial flows would not add up to the gross flows and that, therefore, the net official flows had to fill in?

  • That was just imposed on the chart.

  • There were some assumptions made about continuing certain behaviors and extrapolating them and so forth, but I deliberately didn’t want the red to have to go away, so I told the people who prepared the chart that it should have some red on it. [Laughter]

  • So the red is there first, and private financial inflows are the residual?

  • Well, these are hypothetical. Of all the pieces of the story, the composition of the capital flows that will finance the U.S. current account is the most subject to change and the least predictable. So it would be foolhardy for us to claim in any sense that we have written down a serious forecast. I simply wanted to show how the elements might interplay and what the pieces would be.

  • It’s actually very helpful to see a base because you can always make your own adjustments. It’s the orders of magnitude that I find really very useful.

  • Similarly, I didn’t want the tan portion to grow too much or to grow too little either. It is important to realize that not only do we have to finance the U.S. current account deficit but we have to take into account the reality of U.S. investors buying claims on the rest of the world. The elements of globalization that are making foreign investment attractive—that are allowing people to diversify, reducing home bias, and adding to this dispersion to which you referred—apply to U.S. investors every bit as much as they apply to foreign investors. So that’s an additional factor. And the flows must abide by simple accounting conditions—they have to add up—in order to make the process continue.

  • Could we have a hostile takeover of world statistical discrepancy, which would help? [Laughter] The issue that is unknown here, as far as the dispersion of the current account balances is concerned, is how far it can go. Unless the circumstances lead to unfinanceable ratios of net claims against a country relative to its GDP, there is no arithmetical limit to how far that dispersion could run. We are not seeing anything in this presentation per se that suggests a limit.

  • No, there’s just the ultimate final condition that it can’t continue forever.

  • That brings up Herb Stein’s famous quote as usual.

  • It’s really quite a useful data system that you have set up to provide a sense of the orders of magnitude. Unfortunately, I was distracted when you were discussing exactly what the definitions were of the figures in the box in the bottom left-hand corner of exhibit 3. This is a share of the portfolio of what?

  • For each of these countries, it is the portfolios of their residents. In other words, line 1 refers to the portfolios of euro area residents, 85 percent of which are held in domestic euro area securities.

  • Okay, it’s the residents of these countries.

  • Yes. And they hold 5 percent of their portfolios in U.S. securities, leaving 10 percent held in non-U.S. foreign securities.

  • I’m sorry to interrupt, but is that the same measure of foreign holdings of U.S. assets shown under the disorderly adjustment scenarios? Is that the same thing?

  • Well, yes, in the sense that the figures in column 2 would be the same. But the denominator here is the portfolio per country.

  • And the denominator in the other is the country’s own GDP?

  • So even though the euro area has only 5.4 percent of its portfolio in U.S. securities, that figure is equivalent to 37 percent of its own GDP?

  • Yes, but let me amend my answer to you a second ago. In exhibit 3, the figure in the column labeled “U.S. securities” is bonds and equities.

  • And this other one is all-inclusive?

  • This is all claims, including foreign direct investment holdings—claims in the United States through that channel.

  • So European-owned banks are part of it?

  • Yes, and Chrysler and all the other investment assets that Europeans have bought.

  • I’m sorry, I apologize for interrupting.

  • Theoretically, that figure could go well over 100 percent. In other words, you indicated that this struck you as an indication that foreigners are already heavily invested in U.S. assets, but a country like Belgium, say, could end up with gross net claims against other countries.

  • In exhibit 4, my point is that if there is a disorderly adjustment, the balance sheet effects are likely to be felt more by residents of other countries than they will be by residents of the United States.

  • I think that was well demonstrated. I congratulate you. It was an awful lot of work, and although there were no definitive conclusions here, I think you’ve set up a structure that gives us a detailed sense of what is involved in the funding of the international financial system. Questions? Governor Gramlich.

  • First, I tend to be in sympathy with the view expressed by Karen and Joe that this growing external deficit is probably going to cause trouble at some point. It’s hard to know when. But if I understand what you did in exhibit 5, the disorderly adjustment scenarios, it strikes me that you may have over-proved the point. If one looks at this, the real cost of the disorderly adjustment is the decline in foreign GDP. And that, in turn, seems to come about—you have everybody following a Taylor rule, but then the zero bound is a constraint— because countries can’t lower interest rates as much as they have to in order to stabilize their economies. So the question is, What about fiscal policy? Would it be possible for fiscal adjustments, either here or abroad, to stabilize these foreign GDPs?

  • Fiscal policy certainly could help. The models are such that if you want to hit a target, you can raise or lower taxes or raise spending enough to help to some extent. You might not like the other consequences. But the answer to your question is yes. While we did not actually run those types of simulations, the logic of the model is such that fiscal policy could help in that regard.

  • Has there ever been in the international arena any evidence that governments have responded to crises by appropriate use of fiscal policy? [Laughter]

  • And these governments, in particular, are already challenged with respect to fiscal policy. [Laughter] But it is hard to see anywhere on earth, with the possible exception of some Asian countries, where the governments themselves perceive the scope of using fiscal policy in that way, whether they could do it well or not.

  • I am not here to praise fiscal policy. But suppose we have a disorderly adjustment. I think it is interesting, when you strip it all away, to ask, What is the problem with that? If the problem is really that policy either won’t adjust appropriately in the case of fiscal policy or can’t adjust in the case of monetary policy, then that’s fine. That’s interesting to know. I think that is what you are saying.

  • We take that point very seriously. And it may be one reason—maybe some of you have noticed it, too—that, when we go to international meetings, it is always the other people around the table who are complaining about the U.S. deficit and worry that it is going to cause a problem. That may be because, in fact, the problems from the point of view of the United States seem quite manageable relative to the picture we get of trying to figure out how this situation is going to work itself out in the rest of the world.

  • Governor Gramlich, may I add that I agree with the outcomes of the scenarios, but there is also a lot of redistribution across industries going on in the background here. There’s a shift from capital goods sectors toward export-oriented sectors and the ones that were competing with imports. I don’t know to what extent the costs associated with those types of transitions are built in, but that could be an extra drag, at least for a while, that works itself through the economy.

  • As long as I have the floor, could I ask Linda a question as well?

  • Of course. I don’t know if you’re going to get an answer! [Laughter]

  • Well, I’m sure I’ll get an answer. Actually, I’m not sure I have totally grasped all this. In looking at your exhibit 1, suppose the dollar changes a lot. There are three ramifications on industry that you’ve highlighted. One is that the costs go up, and that’s bad; that’s illustrated by the yellow bars. The import share is high, but it could come down if the dollar goes down, right? And the black bars, the export share, could go up. So you have three things going on for each of these industries, and unless I missed it, I don’t see where you have put it all together. Have you done a simulation with a given percent change in the dollar and looked at the cost structure, the import competitiveness, the export competitiveness, and so forth, to see how these industries come out in total equilibrium?

  • What I’ve done is a bit more limited in that I’ve looked historically at the effects of exchange rate movements on profitability and the net effects in terms of the first two bars, which are the revenue exposures versus their cost exposures. And I’ve found that those effects are a very significant issue for them, as is initially this point about the translation effects on multinational profits.

    The other thing I’ve looked at, where there is industry-level detail, is the sensitivity of exports and imports to exchange rate changes. Those are the two pieces. What we see in that exercise is that, when the dollar depreciates, the exports of these types of industries respond significantly, so revenues get a really big boost. On the import side, though, for these industries it’s a different story. If you look at the overall U.S. economy and at how much we are spending, say, on imports of transportation goods from abroad, you find that the sensitivity of that to exchange rates is pretty much zero. And the reason is that we end up having higher prices on imported cars, for example, and importing fewer cars, so the balance of those effects won’t show up in our import balances. So import sensitivity, the amount we spend on imports, is a different story from the one conveyed by the production volume in these industries since the U.S. producers will still be expanding their own activities.

  • Incidentally, the input–output cost data you have are now for ’92 to ’97?

  • Were these numbers in exhibit 1 for a specific year?

  • Yes. These numbers are, I believe, for 2002.

  • You’ve estimated the costs using the 1997 input–output data, which is close enough or should be.

  • Just a quick follow-up on exhibit 5 and the disorderly adjustment scenarios. You indicate that for the United States you have a Taylor rule kind of reaction for policy. In the most benign case, scenario number 1, what kind of adjustment do we need domestically and—to the extent you can characterize the international adjustment— internationally to respond to this scenario?

  • In scenario 1, we have the federal funds rate following a Taylor rule. The funds rate increases 70 basis points in the first half and gradually moves up about 200 basis points. I don’t have a weighted foreign adjustment, but I can tell you that, in the case of our major trading partners, they hit the zero bound in that scenario.

  • Even in scenario number 1?

  • Yes, but just barely. In scenario 2 they are quite strongly constrained; in scenario 1 they just barely hit the zero-bound constraint. I could tell you how much they would have to lower rates if there were no zero bound: Japanese rates would be down 300 basis points, and euro rates would be down 400 basis points.

  • If they didn’t hit the zero bound, your scenario 1 line wouldn’t go much below zero, would it?

  • No, because they just barely hit the zero bound. So, it’s true that if you’re looking at the effect on their GDP—is that what you’re looking at?

  • In that case, in the bottom left panel, the dotted black line goes down to minus 1½ percent rather than minus 2½ percent.

  • In some sense, if monetary policy could work immediately and without friction, none of the GDP lines would move in this scenario 1.

  • Just a follow-up. The orderly and disorderly adjustment scenarios are based on where we are today, if I understand correctly?

  • Yes. And in some sense, the elephant in the room is the size of the current trade balance.

  • You’re saying that this can continue, though you don’t know for how long, and as it does—

  • The elephant grows.

  • The consequences grow. And yet there is nothing we can do about it, which is an interesting conclusion.

  • In some sense, there’s nothing we need to do about it. On the other hand, I do lose a little sleep over that conclusion. But I do believe that global capital markets are better at deciding where the world’s capital ought to be put than almost any other mechanism I can think of.

  • But we know that at times it gives us a terrible headache.

  • It can. And it’s certainly true that factors like U.S. fiscal policy bear on the intertemporal consequences of allowing this deficit to run. It bears on a much broader issue of intertemporal resource utilization in this country that includes our borrowings from abroad and our repayments or servicing of those borrowings going forward. So, I am inclined to say when I look at exhibit 2—I agree with the Chairman—that it could happen. It is perfectly plausible that there will be no adjustment through 2010. But then I look at the diagram on the top left in exhibit 4, and every year that goes by that the deficit gets bigger, that line shifts. So what it would take to get us going in the other direction becomes a bigger and bigger hurdle over which we have to jump. The interplay of those things and whether the world’s economic resources are being best allocated by all the millions of decisions that are being made by people everywhere is the conundrum in all of this. It is not an immediate issue. It is not going to produce a crisis tomorrow. It might not ever produce a crisis. But embedded in this situation are these issues that are a concern.

  • We have a long list of people who wish to speak, but I think we can break for coffee at this stage. Let’s get back in ten minutes, plus or minus thirty- three seconds.

  • [Coffee break]

  • Thank you. I’d like to take us back to a point that Linda made in passing because it’s one that worries me a bit. There is the sense not just that an adjustment between the consumption of foreign-produced goods and the consumption of domestic goods is involved but also that there would be a switch in the nature of goods, with a move toward fewer investment goods and perhaps more final consumption goods. As you’ve looked at this— particularly you, Karen—have you thought very much about what the implications would be for the U.S. economy before we get to the disorderly type of situation—in an even less benign case than described in the list of implications in exhibit 4? I suspect one of them would be less investment and potentially slower productivity growth here over the long term. Linda, maybe you could start by picking up on the points you were making just in passing to Governor Gramlich about investment goods versus consumption goods. How might that adjustment play into it, and more broadly, what are the implications of that?

  • Well, if the dollar depreciation is part of the larger switch of investor sentiment, for example, out of asset holdings in the United States, we’d have higher interest rates and lower investment spending. In terms of the overall GDP effect, that is offset to a large degree in the simulations by the expansion of production in export-oriented and import- competing sectors. I don’t think the scenarios get into how costly the period of dislocation is— involving the transfer of workers between industries or types of jobs or perhaps a temporary period of capital reallocations. I think that is the context.

  • Linda is referring primarily, at least in the last portion of her answer to you, to what we might call adjustment costs. The mix of U.S. GDP that would come out of this depends entirely on just where the shock starts, how it makes its way through the economy, and in particular, how monetary policy has to respond. In the event that the shock began because asset preferences turned against dollar assets, we would get one sequence of events. We would get stimulus. The United States in any scenario has to produce more tradable goods and fewer nontraded goods. So, in some sense, the shift has to involve a move into manufacturing and certain sectors that we all know and love, and against certain sectors that we may think we probably are overindulging ourselves in anyway. [Laughter]

  • We’ve got a budding politician here! [Laughter]

  • That mix shift doesn’t necessarily have to mean that it is somehow an unwelcome development in the U.S. economy and that we’re going to have a smaller manufacturing sector. We may have a bigger manufacturing sector. We may have more investment in certain sectors, but we’ll have less investment in other sectors.

    Looking at the same set of relationships—and using the vocabulary of the national income accounts—we do have to shrink our savings–investment imbalance. We have to lower total investment and raise total savings so that they meet, or at least come closer to meeting than they do now. The lower dollar will be a terms-of-trade change to the United States. U.S. consumers will have less purchasing power over the world’s goods, broadly defined, than before the dollar fall; that will be a force against consumption, and it favors investment. The really critical thing would be that monetary policy, even in some orderly scenarios I could imagine, would have to tighten to restrain excess aggregate demand that would otherwise emerge in the United States owing to the expansionary shock of dollar depreciation.

    But if adjustment were to begin by fiscal tightening in the United States—if that were among the first things to happen—or by a spontaneous increase in U.S. saving rates owing to demographic effects or some such development that brought down interest rates and exchange rates and lowered domestic demand, monetary policy might even have to ease. In that world the mix of investment and consumption and government spending that we’d ultimately get would be different. So, that’s not predetermined. The mix could come out a variety of ways. But the savings and investment imbalance has to close. That much we know for sure.

  • I would just add, though, that in the case of this asset preference shock that Karen noted, we are going to have higher real interest rates, which will crowd out a certain amount of interest-sensitive spending. Some of that will fall on housing investment and some on consumer durables, but some of it will fall on business fixed investment as well.

  • A lot of this adjustment takes the differential in productivity growth rates in the United States versus foreign economies as exogenous. And what those differentials are really has a great deal to do with the adjustment process.

  • Yes. Indeed, in the background paper we reported another model simulation, which we did not describe today, that makes the shock the productivity story. It involves a one-time productivity shock that works its way through the whole global economy. And you can see how that world looks different from the one in which asset preferences change.

  • But I think to some degree the conclusion from all of this, Karen— your point that exhibit 5 suggests that it’s more the rest of the world’s problem than ours—is not quite right. There’s obviously a reason for all of us to be worried about the outcome from even an orderly adjustment. As you indicated, depending on the way the adjustment starts, we as a society might be relatively unhappy with the outcome or relatively happy with the outcome. So let’s be cautious about saying it’s no wonder that representatives from other countries ask us at various meetings we attend what we’re going to about our external deficit. It’s something we have to be asking ourselves as well.

  • Yes, right. But a stark case of that point is the difference between productivity opportunities emerging abroad that attract capital abroad, as opposed to just a widening of the risk premium because people become disheartened by the course of policy or the politics or the outlook for the United States for some reason. The whole mix of saving and investment globally and the accumulation of capital depends on just these kinds of factors.

  • First, let me compliment everybody who made a presentation today and those who wrote the papers because I think they did an extraordinarily good job—for those of us who don’t live in this arena all the time—of making a lot of complex things quite interesting and quite compelling. I want to ask a couple of questions, and then I just want to make a comment.

    My first question relates to what seems to me a difference between the two papers regarding the impact of a depreciation of the dollar on overall price changes, as measured by the PCE as the Board staff does. It looks to me as though Linda’s presentation suggests an impact that may be a third bigger. So I wondered if I got the numbers right there or not, and I may not have.

    The second question was a reaction to Karen’s statement about why people from other countries are concerned about this. It may be, in terms of portfolio analysis, that it is not a very big allocation to have in the euro area—let’s say, 5.4 percent of the portfolio in the United States. I know that a broader definition of holdings of U.S. assets was used for that calculation than for the ratio of U.S. asset holdings to a country’s own GDP. But I was struck by those latter figures. If you assume that something of a disorderly nature happens relative to the U.S. currency and the U.S. economy in terms of an adjustment on the trade side and 37 percent of a country’s GDP is somehow involved, I can see where that could cause them to be concerned. People in other countries know that adjustments on our side will need to be bigger as time goes by, and the impact on them could get bigger as well. So, it is not surprising to me that they are every bit as concerned about this as they think we ought to be. It worries me that we are not that concerned.

    Finally, I thought your first chart in the external adjustment material was very interesting. The chart on the bottom right-hand side on U.S. saving and investment shows that over time we had a great deal of net domestic investment and then we had the bust in 2000 with the stock market correction and so forth. And we had a substantial amount of net foreign lending. You made the point that, if net saving had continued at the level that it was in 2000, say, the current account deficit wouldn’t be anywhere near as big now as it is. That, of course, corresponds pretty much in time with the swing of the fiscal deficit from something like plus 2½ percent of GDP to minus 4 percent or so of GDP. We recently hosted a conference where everybody was quoting Herb Stein all the time because of the fiscal deficits and trade deficits. Every speaker started off with a quote from Herb Stein. One of the speakers—actually it was Ted Truman— made the point that the twin deficits aren’t twins but they share a lot of DNA, with the DNA being related to the national saving rate.

    It seems to me that we don’t have a lot of options in terms of correcting something that— though it is getting bigger and its consequences are in some sense more dire all the time— ultimately has to come to an end. In my view, one of the ways that policymakers would have the most control over the outcome would be to do something about the fiscal deficit. It was rather interesting to me that you seemed to tiptoe around the fiscal deficit in your paper. You talked about how that could be a correction mechanism, but there was no prescription discussed in the paper. That may be because we don’t control fiscal policy; I suppose that’s possibly why. But that seems to me something that may deserve a little more attention.

  • Let me answer your questions in reverse order, if you don’t mind. There is a difference in some of the conclusions we draw about pass-through, and I think it would be worth airing that a little.

    You might ask, What if we just extended that saving line, wouldn’t everything have been wonderful? But the point of that panel on saving and investment and relating it to the other information on the same exhibit is to try to emphasize the general equilibrium nature of this process. It is true that fiscal policy could help. Fiscal policy is the one textbook answer to doing something in the right direction to resolve this problem. But let’s say that government expenditures were cut for whatever reason—just hypothetically, in a thought process—to match whatever might otherwise have happened and we kept the saving rate sufficient so that the line for saving would have been horizontal. A lot of the cutting of government spending would have been on domestically produced goods, not internationally produced goods. There is no magic direct channel from fiscal policy to traded goods or to our trade deficit.

    We do simulations asking what the outcome for external adjustment would be if we assumed a fiscal policy action of this or that size. If we put a contractionary fiscal policy in place that lowers domestic demand for an unchanged exchange rate and everything else, momentarily we have slack resources in the U.S. economy. The model then tells us that monetary policy should respond and should stimulate the economy to get utilization of resources back to capacity. But the mix of GDP will now be different. By construction, we cut G, government spending. The question is, Are we going to get more exports? Are we going to get more of some other component of domestic demand? And the answer to that question depends importantly on the interest elasticity of the components of domestic demand, the interest elasticity in our model of the exchange rate, and the exchange rate elasticities of exports. So as we do this exercise, filling that hole from fiscal policy, we get back a fairly substantial amount of domestic demand. And the adjustment that we get in the external balance is only a portion—and it could even be a small portion—of the change that was made to fiscal spending that started it.

  • So basically we would need three things to happen domestically and one internationally. On the international side, foreign economies have to grow faster. Domestically, we have to have more saving, hopefully not too much less investment, and the fiscal deficit has to go down a little.

  • Right. I think we’re confusing endogenous and exogenous things here, which makes me a little nervous. But somewhere along the way there needs to be a mechanism that “crowds in,” to use David’s language, exports and helps to get external balance out of the fiscal action. The same story would have been true had the original shock just been a spontaneous increase in saving rates on the part of the private sector. What delivers the external adjustment is that the exchange rate changes; and it changes the relative price among the components of GDP. The result is that people, both producers and consumers, shift their behavior in a way that makes net exports fill that hole as opposed to other components of domestic demand. The interest elastic components also come in to fill that hole.

  • And in the staff model, it would be about $1.00 of trade adjustment for every $3.00 of fiscal contraction.

  • You did have a 30 percent figure in there. But 30 percent is better than nothing.

  • I will take thirty seconds on this question about whether these numbers in exhibit 4 are big or little, and then we will turn to the pass-through question. It is true that the 37 percent of GDP for Europe or the 56 percent of GDP for Canada that are shown in that exhibit are big numbers. But remember, on this basis, capital output ratios for economies are numbers like 300. So, though these economies hold claims on assets that are multiples of their GDP—and their claims on the United States may be 60 percent of their GDP—you want to put it in that perspective.

  • We’ll turn to the pass-through question. Linda.

  • We do have a disagreement on what we think the pass-through is. I’m guessing that, in the end, there won’t be as much disagreement about that going forward as it might appear. Let me go through what the difference is. When I talk about pass-through of exchange rates, I’m referring to the pass-through of exchange rate movements into import prices, and I’m giving as a rule of thumb a pass-through of one-half. In the Board’s simulation model, I think it’s about a third, or 30 percent.

  • It’s about 25 percent in the Board’s simulation model. So we have a different sensitivity of import prices to exchange rate changes, and that has a number of effects in terms of what you see here in our materials. Looking at exhibit 5 in Karen’s and Joe’s material, the disorderly adjustment scenarios, if the pass-through of exchange rate movements into import prices were higher, there would be more import-price response and more core PCE response. We’d probably have as well a larger relative price adjustment between imported and domestic goods, and we would see more of a net export response, more expenditure switching going on in the model. That’s how that change would manifest itself in that scenario; a higher pass-through would give you more inflation and more of a net export response.

    So then the question is, Why do we have a disagreement? Part of it reflects how we estimate this, and part of it reflects how we view the current data relative to history. Is the recent history most relevant? Is the longer history most relevant? What I see is that the pass-through did decline; it seemed to be lower over the 1990s. But recent price observations, while on the low side, are still within historical bounds. So which is it? Do we have a permanently lower relationship, or are we on the low side of something that historically has been high and low? I tend to go with the longer history in part because, after we’ve had a strong dollar period, foreign producers have been able to be quite profitable and build up their margins. That also gives them scope for accepting lower margins for a while, but that’s a finite process.

    One interpretation is that, in 2002 or early 2003, foreign producers were living off the fat of a strong dollar period, and maybe that’s not a long-run relationship to be locked in. Also if you think of the Asian countries, China’s prices haven’t been changing against the dollar because their currency is pegged to the dollar, so China’s competitors might have been more constrained in moving their prices. Whether or not that is something that can persist is debatable.

  • I think we agree on what we’ve seen in the past and, of course, we both are wondering what will happen in the future. I’d just like to say that I believe a lot of the differences in our views come from how we think about commodities. That’s an issue to consider going forward. We both looked at import prices, disaggregated and aggregated. When we look at aggregated import prices, we tend to put commodity prices separately in our regression in addition to exchange rates. We, too, find that there was less pass-through in the 1990s. The question is, What is the story lately? I think that’s where we may have some difference.

    In the past few months we’ve been saying that more of the rise in prices has occurred because commodity prices are rising, which we think has some exchange rate component but also has a large cyclical component—domestic demands in China or, more broadly, a world demand component. To our mind, that may explain why the pass-through seems to be higher lately. Time will tell which will turn out to be right. Frankly, I have a lot of sympathy with perhaps a return to a higher pass-through in the future because I think, as Linda said, the lower pass-through in the ’90s might have been because it was a time when China’s economic growth was rising and their currency was pegged to the dollar. Even though a lot of countries in Asia had exchange rate movements up and down in the ’90s, for those who compete with and integrate their production processes with the Chinese, there was an 800-pound gorilla that they couldn’t really compete against. So they were forced to price along with China, and China’s currency was tied to the dollar and thus our pass-through estimates would fall. When we’ve looked bilaterally at which countries seemed to have passed through less in the 1990s, it was Japan, developing Asia, and Mexico. So it’s quite possible that, if China were to revalue or think of a change in that respect, all bets could be off. Maybe we’d go back to the historical pass- through. The jury is out on that.

  • Mr. Chairman, my question was asked already, so I’m going to pass.

  • Thank you. Joe, in modeling the current account deficit, you have to explain where it came from. A big part of your story is that U.S. assets and foreign assets are imperfect substitutes, and there was a big exogenous increase in the foreign demand for U.S. assets. I understand why you did that, but that’s a very important assumption. The elasticity of substitution between our assets and foreign assets is going to affect many aspects of the model. For example, if it’s inelastic, a change in the demand for U.S. assets on the part of, say, foreign central banks would have a large effect on the value of the dollar. So I guess my question is, Do you accept that implication? What would be the effect of assuming a much more elastic relationship in demands between different assets? How central is this to your basic story?

  • You are referring to the background paper that we circulated?

  • In that we had a somewhat different model simulate a rise and fall in the current account balance. I’m not sure that we are here to hang our hat on one explanation for the rise and fall of the current account; clearly a number of things were going on then. I wouldn’t say that the portfolio preference shock is entirely the story. But one reason that I chose to highlight it, to some extent, was that, if you look at a number of variables in the U.S. economy at the time—interest rates, output, inflation, the exchange rate, the trade balance, and so forth— it’s the one shock that sort of moves things jointly together well. That doesn’t mean that some other combination of shocks might not also have worked out that way. But it is one that does work out that way. So in that regard it has some nice features. Another obvious shock that we highlighted was the productivity shock, and the reason we didn’t use that alone is that it just doesn’t do enough. You’d need a much bigger productivity shock in our model for any reasonable calibration to explain the trade balance.

  • I was also pointing out that this assumption has important implications for the way the model works. For example, relatively small changes of demands in the portfolio can have large effects on relative asset prices. How sensitive are your simulations to that assumption?

  • That’s a very good question. I don’t sense that it is that important. A movement in foreign exchange intervention of, say, $100 billion or so in the model is not going to swamp the exchange rate, if that’s what you’re saying. Is the portfolio balance built to do that? No, I don’t think it would do that. I actually can’t point to specific evidence, though, so maybe we should look into that. But my sense is no—that the simulations are not sensitive to that assumption—because these are very large numbers. Even intervention of large amounts by the standards we are used to is not going to be large enough to make a difference. We’re talking about a trillion dollars. But your point is a good one. As a modeling convention, it really helps in terms of actually solving these models. It’s quite normal in these kinds of models to put in something like this because the model can be stable and you get to a steady state you can understand. So it’s not unusual to do that, but it’s a good point.

  • Coming back to the point that several have made about the frustration with the absence of a response to the external imbalance issue, there has been talk about a fiscal policy response and even a monetary response in the instance of a disorderly threat. My question—and I asked this of Joe at the break—is what the effect would be if the United States signaled a change away from a strong dollar policy. That is a posturing change, assuming we are not actually intervening. What would be the impact of the Secretary of the Treasury saying, à la John Connally, “We’ve thought this through again, and actually a 10 percent drop in the value of the dollar would seem to be appropriate at this time”?

  • The one stumbling block in trying to analyze this problem that we’ve encountered since we began—we have a little cottage industry here, and I have been doing this for eight or nine years of my life now—is that we don’t have the capacity to relate exchange rates, in real time or hypothetically in a model, to a number of variables and have a deterministic relationship among those variables and exchange rates that works very well at all. So we are always forced to pose the problem and structure the answer in ways that don’t require that we claim that we can forecast exchange rate developments. We talk about relationships, but we don’t say that the dollar will do this at a certain point in time or even that the dollar will do this in response to the current account, which is an open issue that we didn’t even raise here. We have lots in here about how the dollar affects the deficit implicitly. In Joe’s simulation, the dollar is driving these things. We don’t really have anything about how the deficit is feeding back on the dollar because being able to explain determination of exchange rates is the weakest link in international economics going. We don’t have an answer, and others don’t have one either. I’m inclined to think that, if Secretary Snow rented out Yankee Stadium and made the announcement you indicated, the dollar would move. But I wouldn’t be basing that on an understanding of economics. [Laughter] It would be more my understanding of the politics in the United States and what I think traders would do because that’s what they think other traders would do.

  • Well, let me come at it from the other direction. What is the impact of staying pretty close to the position generally espoused by the Secretary of the Treasury—except for O’Neill who was somewhat more flexible on it—of being supportive of a strong dollar position? The current Secretary has maintained that position, at least as I understand it, of being supportive of a strong dollar.

  • Part of me clings to the hope that oral intervention of whatever sort has a marginal and only transitory effect on exchange rates and that deep down the economic fundamentals drive the process, regardless of what anybody says.

  • So that ought to work the other way also, on the downside?

  • I have a question for Karen and Joe. Karen, do you want to hazard a view on whether the probability of a benign adjustment exceeds that of a malign disorderly adjustment?

  • Ask her to predict the exchange rate!

  • We’re just not in the business of predicting crises here. Down deep the things that make the U.S. economy different from the rest of the world I think are overpowering in terms of the role of the dollar, notably the size of the economy, the flexibility of the markets, and the depth and size of the asset markets. The notion that the world’s investors will truly become so disheartened about the U.S. economy that they will produce something that would look like a disorderly adjustment strikes me as remote. On the other hand, I was an adult in the 1970s and early 1980s, and I remember periods when the dollar was thought to be a very weak currency. And I remember periods when Federal Reserve policies weren’t well regarded. So I’m not about to say that could never happen. I just don’t see it happening any time soon.

  • So in a sense, you think we control the risk?

  • If U.S. macroeconomic policy remains well grounded, if the value of goods in terms of U.S. dollars is well maintained by this Committee, if nature doesn’t impose some truly exogenous crazy event on the U.S. economy, and if U.S. fiscal policy could get a little better, I think those are the things that will determine how this situation evolves. But, you know, when I read my own briefing I realize what it means—that this problem is going to continue to grow. And I remember Herb Stein’s admonition and say that this situation can’t go on; at some point, in some way, an adjustment has to occur. The paper had a rather benign outcome in which the productivity differential just worked its way through the system and things more or less corrected. I can certainly imagine, if the same sort of productivity shock that hit the United States were to become evident abroad—which we keep looking for—that it would accelerate that process even more. And those stories do exist. Those adjustment paths are very credible.

  • I want to come back to scenario 1 in your exhibit 5, in which you have the dollar falling 30 percent in a year. Certainly an outright cliff of that magnitude in a year—

  • It’s in two quarters actually.

  • In two quarters. And you assume that comes with no increase in risk premiums on other U.S. financial assets. I’m asking how realistic is that?

  • No, no. You should think of scenario 1 as just an initial description of an element that is in the whole set of scenarios. We didn’t really mean for you to take scenario 1 seriously without scenario 2 coming along as well. That was more an analytical clarification. We think that much dollar depreciation would roil markets.

  • In that short a time period.

  • Yes, in that short a time period. As indicated in exhibit 6, starting in 1985 a depreciation of that magnitude occurred over 2½ years.

  • And in 1986 we sat here and had a conversation—one in which I actually participated, sort of in Linda’s role—and tried to ferret out what the consequences were going to be of the dollar depreciation that was in train and how much dislocation there would be. We were all concerned about whether there would be enough domestic capacity to meet the shift in demand that was going to take place toward tradable goods and away from other goods. We wrung our hands a fair amount about that. But in retrospect, it’s clear that there’s a lot of flexibility in the U.S. economy. And given that the shock took place over 2½ years, there was in a sense plenty of opportunity, therefore, for the private economy to respond more smoothly to it. So there was less risk associated with the rising risk premium in financial markets. That process occurred more smoothly, I think, than we thought was likely to be the case ahead of time.

  • In my own mind—Joe actually did this page, so he might have a different way of explaining it—I don’t think of the shock in scenario 1 as causing the shock of scenario 2. If people were to wake up and decide that they didn’t want to hold claims in the United States, that would mean that all those things would occur at once. People simultaneously do not want to hold dollar claims but also do not want U.S. equity or U.S. bonds, et cetera.

  • So you’ve drawn no particular policy implications in terms of things we could do to mitigate the risk of an adverse outcome beyond what we would do anyway, such as try to keep monetary policy credible, inflation expectations low, and rediscover fiscal virtue. Another question is, Does it matter at all what happens to exchange rate policies in greater Asia?

  • I see that as affecting the path. In other words, those countries are in some sense financing our deficit. Official finance is a variant of private finance; it has different incentives behind it and perhaps works through other asset markets a bit differently. But what the Asian official sector is now doing is in part financing a continuation of the deficit as opposed to triggering or even participating in the adjustment process. There are some risks embedded in that. There are some who feel that the decisionmaking in the official sector is a bit more political; it is not comparable to the kinds of decisions and thought processes made by those in the private sector. Also, the sums are getting very large so that the holdings of foreign claims on the United States are becoming concentrated in a very few spots. Huge sums are held in the Central Bank of Taiwan and the People’s Bank of China; that in and of itself sort of changes things a bit. On the other hand, I’m inclined to think that those official entities are not likely to shoot themselves in the foot if they can help it. So there are pluses and minuses, I guess. If they change their exchange rate strategy, I would expect them to change their reserve management strategy; I think those two things will go hand in hand. While it will matter to the specifics of the path, I don’t think it’s a deal breaker.

  • May I just add one point? I fully agree, but in addition to the exchange rate policies in greater Asia, there’s the question of whether demand management there can ultimately help push toward the benign outcome. If we get increased openness and thus greater consumption of U.S. goods—so that basically those markets become more of an export destination for us than they have been—that can help facilitate the export expansion that would produce a more satisfactory outcome. That said, Asia is still a pretty small share of overall world spending and that effect isn’t going to be tremendous; but it could help on the margin.

  • Actually, I was going to ask about this last point, the Asian dollar block. Let me just ask one variant on that. The Europeans sometimes complain that the Asian dollar block forces more adjustment through them. But I would say that that is not necessarily the case. It could mean that the Asians, by absorbing these claims, are stopping the total adjustment from happening. It’s not necessarily true that they—

  • That is certainly the attitude that the staff has taken. We’ve even written those points down, and we’ve asked the Chairman to say them out loud, and he has. [Laughter]

  • And nobody knocked me down!

  • You’re just following orders, I know. [Laughter]

  • In part, it goes back to this unanswered question of exchange rate determination. In essence, that conclusion follows from a kind of portfolio balance theory of the exchange rate. What it says is that, when the Chinese, for example, take dollars off the market and out of the hands of the private sector, they make dollars less abundant relative to Chinese renminbi. That is true, but they also make dollars less abundant relative to European currencies. So in essence, they absorb some of the dollars that would otherwise be putting pressure on the euro and any other currency you care to name. That’s because there is this big portfolio balance issue and relative supplies are a factor; so anybody who buys dollars in some sense is helping the other currencies, not hurting them. But that isn’t the only theory of exchange rate determination that one might want to adhere to. And given that none of these theories is empirically very satisfactory, it is possible to construct the story another way and get a different conclusion.

  • While I have the floor, let me make one comment on fiscal policy and the exchange rate correction—the fiscal rectitude point. I’ve told the same story that we’re telling here, but I also remember the lively discussion that Chairman Greenspan started in Jackson Hole about ten years ago, when he said that he thought narrowing the federal deficit would strengthen the dollar. I think the channel he was thinking of was confidence. So if we had a circumstance where the fiscal situation looked explosive and we started narrowing the deficit, it’s very hard to tell what the effect on the dollar would be. The models say one thing, but who knows in real life?

  • There are time-varying and policy-varying risk premiums on most of those equations, and we can get different outcomes.

  • I have two questions. The first one is for Joe and Karen because I was thinking about financial markets as they apply to the analysis that you did, especially in the first paper that was distributed to us on June 16. When I look around the world, the United States has really been very innovative compared with other countries in terms of the types of private securities traded in our markets in part because of the depth and liquidity of our markets. We have been innovative in securitizing duration and credit risk—and in every other way that one can cut and slice risk—and in creating derivatives, too. My question is whether in your analysis of the private securities flows you see something of a supply-side impact on those flows. If someone is sitting in another country doing good portfolio analysis, and the depth, liquidity, and variety of risk exposures they can acquire in U.S. securities is more attractive than the alternatives, is that in and of itself a tradable good that attracts them to the U.S. markets?

  • I would say it’s not. The fact of the matter is that most of the securities of at least private corporations—not U.S. Treasuries, obviously—that foreigners buy are marketed in Europe and are sold in the euro markets. They are not sold in New York. I’m not saying that there aren’t any sales of the type you asked about, but there’s a huge offshore market for corporates in which the securities are denominated in various currencies and the lenders can vary. And that market is for the most part in London—in Tokyo, too, and to a limited extent in Frankfurt—not New York. What foreign buyers of U.S. corporates purchase is a security issued in their own jurisdiction and in some sense governed by their own jurisdiction’s practices. So, while I see your point and I’m sure that some part of what you are saying is accurate, it’s not as if these people truly are participating directly in the U.S. corporate market. The U.S. corporations are going overseas to the euro market.

    What is true is that foreign corporations haven’t followed those in the United States in choosing to offer corporate tradable securities that have a range of alternatives. The secondary markets haven’t developed, and the derivatives markets haven’t developed as much, in part because of the universal banking tradition that they have. They’ve gone a different way. It’s true that they might change. Those corporations could become more multinational so that their taste for how to finance themselves in different places changes. If European corporations start to offer a richer mix of high quality corporate debt in their own markets, European or Asian or other non-U.S. investors will see more competition with U.S. bonds and equities, and that could have an effect.

  • My other question is for Linda and goes back again to exhibit 7 and is about the pass-through. One of the changes that international corporations have made involves replacing trade by going into customer markets—actually putting in facilities and manufacturing in the resident country where the ultimate sale occurs. When I look at the data in exhibit 7, the industry where the pass-through has come down the most is transport, which I guess is composed primarily of autos and small trucks. Clearly, we’ve seen German and Japanese auto manufacturers moving into the United States more of their production of vehicles destined for this country. By putting more of their production in the resident country, does that in effect reduce their inherent exposure to foreign exchange risk? Is the amount that is still being imported into this country small enough that they don’t need to push it through as they might have done twenty years ago, when all of their U.S sales were being imported?

  • I think that’s certainly a possibility. When I think about what the import-price response to an exchange rate change is going to be, I think of the producer who is doing the exporting looking at his own costs and his margins over those costs. Based on that and the markups in the industry and so forth, he then determines how much he has to change prices. So that part doesn’t change even if the company has facilities in the United States or has diversified the locations of its production. There’s another part to your question about whether production diversification across markets is in part induced by exchange rate volatility, and people really differ in their view of the importance of that. I tend to think that diversification occurs in part because firms want to offset some of the risk by having the ability to move operations between markets. So I agree with that, but I don’t know how big that is relative to the total amount of investment and the total amount of trade that goes on.

  • First, let me add to Cathy’s commendations for Joe and Linda and the excellent background papers. My question relates to Tim’s in a way. It has to do with the locus of disorderliness. The prospects for disorderly adjustment seem to motivate and animate a lot of this analysis. And it strikes me that a lot of the risk regards the rest of the world. Some of our trading partners have been known to resist the appreciation that accompanies the type of international adjustment that we’re looking at here. In particular, some developing countries often resort to macroeconomic commitments like exchange rate pegs that they have only a limited ability to follow through on in a consistent way. These commitments are not perfect substitutes for the credibility that comes with deep and resilient markets and a credible commitment to low inflation. It strikes me that those kinds of commitments by these countries enhance the prospects of their being susceptible to domestic macroeconomic instability in response to changes in U.S. macroeconomic conditions and the type of adjustment that we’re looking at here. I wonder if you’ve given any thought to that or what your general reactions to that notion are.

  • Well, a certain variation of that theme is to rerun the Asian crisis in one’s head over and over and ask, What if this or that or the other thing had been different? Indeed, many people now argue that some of the Asian countries are recreating conditions comparable to those that existed in 1996 and 1997, in particular the stability that they are trying to achieve in their exchange rates. Are those kinds of conditions being recreated? I think there is a clear risk.

    Now, it is also true that one of the lessons that those countries drew from the Asian crisis is that large holdings of reserves are a good insurance policy. Also, they realize that running a current account surplus removes the need to go to the market to finance their activities in any given month, day, or quarter, if conditions get a little adverse. I think they are quite explicitly seeking to build large quantities of reserves both to impart confidence to the market, which sees that they have all these wonderful reserves, and to use if needed—though I would say it’s probably more the former than the latter. And they exhibit a certain bias in the direction of keeping their exchange rate from appreciating in real terms because that’s what generates the surpluses that protect them from the need to go to the market. So they don’t have a large outstanding stock of debt; and, if anything, they are running down their net debt position and might have positive claims on the rest of the world at the rate they are going. They don’t have a flow problem, and thus they feel better positioned to manage the ups and downs of being globalized into the world’s capital markets.

    The problem they face involves the consequences for their domestic economies if these strategies ultimately get uncomfortable. One thing that could happen through the back door is that they may end up with their money supplies growing and they may have domestic inflation; so they get real exchange rate appreciation even while they are maintaining nominal exchange rate stability. They may at the same time lessen the chance to strengthen and develop their banking systems and their capital markets because they are constraining prices in the domestic financial system more than is attractive. So they may be encouraging their private citizens to diversify out of the domestic market, even though they are trying to reassure those private citizens through the exchange rate stability. These countries might be better served in the long run not just to hold on to the resources they are requiring as reserves but to use them to come up with a stronger banking system and more competitive capital markets. Ultimately that strategy of trying to achieve exchange rate stability is not going to prove to be a long-run winner. It could be a good adjustment strategy, buying time to accomplish the domestic reforms and to strengthen their financial systems—doing that carefully over a period of time. But it does have the capacity to put us back in a world in which the emerging-market countries, even successful ones, lack some of the features that the developed countries have, particularly on the investment side. So these emerging-market countries are still subject to concerns about sudden reversals of capital flows or capital flight from their own residents, and those sorts of things. Indonesia, for example, has had a lot of pressure on its exchange rate lately—shades of a return to 1997. It’s a real issue. And I think it’s one of the reasons that we can finance a 5 percent deficit when they probably can’t and that we can probably finance a 6 or 7 percent deficit when they surely couldn’t. Does it mean that we can finance a 9 or 10 or 15 percent deficit? I don’t know.

  • I know it’s about time to end this discussion, but I have one more question. Does it matter what happens to the projected foreign share of the U.S. Treasury market under a scenario where our external imbalance is substantially larger—maybe getting larger than the increase in the net borrowing requirement of the government? Is there a point at which you would worry about the implications of these lines going on indefinitely, with a growing foreign share of the outstanding stock of U.S. risk-free assets?

  • Within the kinds of numbers we’ve experienced to date we’re inclined to be relatively relaxed about that. Dino might want to comment on that issue, too, though.

  • Doesn’t it depend on the interaction between U.S. Treasury securities and private securities in total?

  • You assume that the preference would change?

  • Yes. In other words, if there is literally zero preference difference, then the 50 percent foreign holdings of U.S. Treasuries is the wrong ratio. The denominator has to be the aggregate. And it’s probably closer to that than not, so that number falls from 50 very significantly.

  • You assume that their preference would change as their share increased?

  • I think what we’re talking about are the preferences of everybody else. Let’s assume that foreign officials, because of laws written into their books, their regulations, their procedures, or whatever, always want U.S. sovereigns but that the vast majority of global investors regard U.S. sovereigns and U.S. corporate triple-As as perfect substitutes. Then this tiny group called foreign officials can do whatever they want, and it really doesn’t matter to the pricing of the assets involved.

  • And if the yield spreads are relatively stable, then the substitutability, for all practical purposes, is without limit.

  • Now as a taxpayer, Mr. Chairman, it would probably be churlish to complain that the foreigners are buying our debt at high prices.

  • I agree. It would be churlish. [Laughter]

  • I think there are consequences for market liquidity within the Treasury market because some of these foreign purchasers do act differently in the primary and secondary markets than private holders. In particular, there are differences in terms of participating directly in the auction and not putting in competitive bids and in the secondary market not putting the securities out to lend.

  • Aren’t these second-order effects generally?

  • This reminds me of the debate we had when we thought U.S. debt was going to disappear. While that has consequences, it was not something we regarded as a bad thing because we had surpluses. So, surely it is manageable.

  • We started a whole new discussion! [Laughter] Anyway, the clock has run out on us, and I just want to congratulate Karen, Joe, and Linda. It has been a very interesting seminar, which is really what this has been, and I hope it engages us. One little thing that went by unnoticed as you were talking about “nature’s” effect or an exogenous event: Is Osama bin Ladin part of nature?

  • I say that because there is a very interesting phenomenon here, in terms of what happens to all of this adjustment process in the event of a significant terrorist attack within the borders of the United States. That’s a whole different scenario, which we really haven’t addressed in any material way, except to presume that, if we can manage a high degree of flexibility in our markets, it will somehow get absorbed. But how it will be absorbed, I’m not sure we know.

    In any event, as you are all aware, we have our annual reception at the U.K. Embassy this evening. Cocktails are at 7:30, and usually we sit down for dinner about a half-hour later. Transportation has been made available for Board members already. For the Presidents staying at the Watergate, vehicles will be available at your hotel at 7:15.

  • Mr. Chairman, I was wondering if there would be enough time for Dino to give his report from the Desk. That way we can start tomorrow with the chart show.

  • I’m sorry. I didn’t realize that we did have some time. I think that’s an excellent suggestion. Dino, why don’t you get started and see if you can finish.

  • Thank you, Mr. Chairman. I’ll be referring to the charts that Carol circulated a short time ago. I’ll try to proceed quickly. During the intermeeting period the market’s focus was on the anticipated start of the tightening cycle. As shown in the top panel on page 1, three-month deposit rates (the black line) increased nearly 50 basis points and forward rates three, six, and nine months forward (the dashed red lines) rose more than 50 basis points. The nine-month forward rate now is above 3 percent. Treasury yields moved higher. The middle panel graphs the two-year yield since January 2002 along with the target fed funds rate. Two-year yields have nearly doubled since late March to about 2.85 percent, and their spread to the funds rate is at its highest since the spring of 2002. Much of the focus in markets has been on recent inflation reports, and the short end of the curve has responded. The long end has not risen as much. The coupon curve, represented by the spread between two-year and ten-year Treasury yields, has flattened to below 190 basis points, as shown in the bottom panel. One thing that market participants have pointed out is that, if expectations were for sharply higher inflation, one would expect to see that curve steepen, but it has not.

    Another way to look at inflationary expectations is through the TIPS market, though here the usual health warnings about interpretation of TIPS yields and breakevens apply. As shown in the top panel on page 2, breakeven yields have tended to rise in recent months as the economic data have improved, oil prices have stayed elevated, and the CPI index has risen. A noteworthy feature of recent trends is the convergence of breakeven rates between shorter-dated TIPS and the longer ten-year TIPS. One possible interpretation is that inflationary expectations in the near term have risen.

    While that could be a contributing factor, there are technical factors that may be having a disproportionate effect on these relationships. TIPS are tied to the not- seasonally-adjusted headline CPI rate, which does create some interesting dynamics in the pricing of the TIPS themselves. The recent rise of oil prices has increased investor demand for TIPS because of the anticipated additional inflation accrual on TIPS principal. This may not reflect expectations for broader inflation to take hold. A second technical reason is related to the calculation of the inflation accretion; shorter-dated TIPS outperform longer-dated ones during a period when the CPI index is rising. This will reduce the real yields and, in turn, widen the breakeven. Third, the TIPS market is beginning to mature. Pension and investment consultants have discovered TIPS and are increasingly viewing them as a separate investment class. The movement of funds into TIPS from that source may be exaggerating the price movements and lowering real yields for the time being. We don’t know how much money has been allocated to this asset class. But we can look at what is happening with regard to the dealers. In the early years of the market, the dealers tended to be long TIPS. One has to assume this was voluntary on their part and that they found the yield attractive. This is depicted in the middle panel. In recent months inventories have been near zero, or net short TIPS, presumably as the real return has become less attractive to this group of informed agents.

    With such technical factors making interpretation of TIPS difficult, another place to look at is in one-year forward inflation rates derived from CPI swaps. CPI swaps are an exchange of fixed for floating payments based on the consumer price index. As shown in the bottom panel, the forward inflation curve has shifted upward—in a largely parallel fashion—by about 25 basis points since early April, which is supportive of the notion that there has not been a disproportionate shift in the level of concern about near-term inflation performance.

    Looking at fixed-income markets more broadly, there has been a lot of comment about the degree to which positions have been adjusted in advance of the expected tightening. Or put differently, the issue is, How much of the carry trade has been closed out? Unfortunately the data to answer this question are not available, and the information that is available is ambiguous.

    The top panel on page 3 graphs primary dealer position data since May 2003. Dealers typically are long credit products that are in turn hedged by being short Treasuries. And typically dealers, as a group, are somewhat net long. In the past few months, dealers have extended their short Treasury positions (the dark line), and their overall positions are close to flat. This is suggestive of one key constituency taking steps to reduce risks ahead of an anticipated tightening. Another perspective is provided by position data made available by the CFTC, which are shown in the middle panel. While these data have their own definitional problems, they suggest that noncommercial users were going short Treasury futures, presumably either to hedge other positions or to extend speculative short positions. This is also consistent with an unwinding of leverage. Other bits of evidence are not as definitive. Bankers suggest that hedge funds are still able to secure very favorable trading terms. And while individual hedge funds are less leveraged than they were several years ago, there are more of them around and it’s not at all clear that their strategies are so different—notwithstanding assurances from some about the heterogeneity of those strategies. In addition, sectors of the fixed- income market where the carry trade was said to be crowded have shown only a modest widening of spreads. As shown in the bottom left panel, the investment-grade corporate and MBS spreads are somewhat wider, but the price action does not suggest large-scale liquidations to date, and anecdotal reports suggest that banks and hedge funds are still holding on to large positions in mortgage-backed securities.

    The bottom right panel graphs the EMBI+ and high-yield spreads. Interestingly, spreads—after widening around the time of the last FOMC meeting and after some negative news out of Brazil—have narrowed to below 400 basis points for high-yield instruments and to below 500 basis points for the EMBI. There are certainly some fundamental explanations for the recent narrowing: improved corporate outlooks, lower oil prices in recent weeks, and some improved emerging-market news. But none of these developments in and of themselves suggests that the leveraged trades have necessarily been liquidated.

    Price action in foreign exchange markets has had traders scratching their heads. The top panel on page 4 shows the dollar’s movements since April 1 against a few major currencies—but this time not including the New Zealand dollar! [Laughter] Despite the sharp movements of interest rates since early April and the shift in expectations, the dollar has barely moved; and since the last FOMC meeting it has tended to fall. Foreign exchange traders have taken that as a bearish sign, which suggests to them that the dollar will fall once the tightening cycle actually begins. Several have noted that the dollar fell from about 1.75 to 1.50 against the DM during the 1994 tightening cycle. Japanese markets have performed well; equity prices have outperformed as the outlook for Japan has brightened. The yen has been under modest upward pressure again, and some in the market are speculating that the Ministry of Finance will return to intervention fairly soon if the yen continues to face upward pressure. The middle left panel shows the recent rise in the ten-year JGB yield. The rise in yields has been generalized across the curve. As can be seen in the middle right panel, which depicts the Japanese yield curve as of yesterday and as of roughly a year ago, the entire yield curve is sharply higher and steeper. While the authorities have been talking yields down the past few days, many in the market are beginning to think about the Bank of Japan’s transition back to a more normal operating structure. In Europe, the data have tended to be mixed and somewhat disappointing given growth in other regions. Reflecting this divergence of performance, yields there have risen only slightly and not as much as comparable U.S. yields—as shown in the bottom panel.

    Turning to page 5, I’d like to spend a few minutes updating the Committee on the reduced volatility in the fed funds market that we have observed in recent years. The type of rate volatility I will be describing is very short term: intraday standard deviations of the funds rate around the daily effective rate and the deviation of the daily effective rate from the funds target. The top panel on page 5 graphs the annual average standard deviation of the funds rate around the effective rate in blue and gives the annual median in red. In general, the standard deviation of the funds rate measured by the annual average was in a range of between 15 and 25 basis points until 1999, and it was somewhat less than that using the median. The spikes in the early 1990s were linked to a variety of factors, such as financial problems in the banking system that created tiering and the elimination of reserve requirements on nontransaction and Eurodollar deposits. The decline in volatility began in earnest in 1999 and has been maintained in subsequent years—with some notable exceptions such as September 11, 2001, and the days after. The bottom panel takes the same data and presents them in a different way; it depicts the medians of rolling ten-day periods of daily standard deviations of the fed funds rate. This graph shows the regular spikes associated with the year-end, which have now all but disappeared, as well as the somewhat higher volatilities in the fall of 1998.

    Page 6 takes a different perspective by looking at the absolute deviations, expressed in basis points, from the fed funds target. This measure suggests that most trading is increasingly very near the target, with very little deviation. The precise reasons for this notable decline in volatility are not clear and my colleagues at the New York Fed are doing some research on this question. No single explanation by itself is compelling, but some combination of the following probably would get us most of the way there. First, among technical factors, was the shift to lagged reserve requirements in 1998 that helped make reserve demand more predictable both for the banks and the Desk. Second, the Desk expanded the number of banks from which it collects daily reserve position data— also to help it better estimate reserve demand. Third, banks have developed better internal control and information systems for managing their reserve positions. And fourth, at the margin, the Desk is probably intervening more often now than it did in the late ’80s or early ’90s, given the explicit rate targeting regime—though this does not explain the continued decline in volatility since 2000.

    But the lurch downward in observed volatility is hard to explain with technical factors alone. Two other factors that probably played a role are the lower absolute value of interest rates and the introduction of the primary credit facility. Lower rates directly limit the absolute amount by which rates could decline; so with the target at 1 percent, the scope of deviation on the downside is only 100 basis points at present. Indirectly, low rates have lifted reserve requirements (by boosting growth in transactions deposits) and have created more room for banks to increase their clearing balance requirements consistent with their use of priced services. And possibly the experience with much lower volatility has affected market expectations to the point that the trend has become self-reinforcing.

    The impact of the primary credit facility is hard to assess at present. It has limited the extent of upward spikes in rates on individual days, as shown on page 7. But since we saw few occasions when rates reached close to 100 basis points above the target in the year or so preceding the creation of the new facility, it’s too early to conclude that the facility’s creation has been a major factor in explaining the reduced level of volatility. At some point the impact of lower rates, including the impact on reserve requirements, will be reversed. When this happens there will be some increase in rate volatility. But many of the other factors mentioned persist, so even if volatility increases, there are reasons to expect that it will not rise again to the levels we saw in the 1990s. Two wild cards that could affect this picture include the possibility that the Congress will permit the Fed to pay interest on reserves and, second, the ways in which the new daylight overdraft rules for GSE payments of principal and interest are implemented. Either of these has the potential to have a major impact on fed funds volatility trends in the future.

    Mr. Chairman, there were no foreign operations in this period. I will need a vote to approve domestic operations.

  • I haven’t looked at the primary dealer data system in recent years, but the net outright positions, as you point out, have essentially come down to zero from a significant plus. What are the missing items in the balance sheet that account for the difference?

  • I’m not sure I fully understand your question.

  • Well, on the asset side, the primary dealers have positions, long and short. There is a net position with a value. What is on the liability side for capital?

  • What we try to do is to look at the position data to see what the longs and the shorts are, excluding certain things such as TIPS, bills, and discount notes. That basically gives us a sense of what securities are there. Now, there will be a big repo book on both sides of the balance sheet, so those have been subtracted as well. So this is one way of giving a picture of the position—

  • The tradable positions.

  • Yes, without cluttering it up with a matched book, which would tend to balloon the balance sheet. That would tell us something about the financing, how this on net is being financed. But the matched book operation is its own business—

  • And you don’t have the derivative positions net in this at all?

  • That’s right.

  • One other thing. On the chart showing Japanese three- month to thirty-year government yields, are these the absolute values and the actual yields? These are not the yields of the one-year maturity ten years out?

  • No, these would be the actual yields.

  • They are the actual yields. That leads me to conclude in looking at the ten-year, twenty-year and thirty-year yields that the more implicit one-year maturity ten years out is yielding very significantly above these numbers. What is that yield, do you know offhand?

  • In our case, we’ve got 6½ plus percent, and this doesn’t look all that different.

  • This goes to thirty years.

  • I know, but look at the ten-year. It’s the rate of change in the yield curve.

  • There’s a risk premium in that as well as an expectation.

  • I would hope so. [Laughter] Questions for Dino? You traumatized everybody!

    SPEAKER(?). I move approval of the domestic transactions.

  • Without objection they are approved. We will adjourn until tomorrow morning. Is anyone not coming tonight to the dinner? Good, I’m glad everyone will be attending. So we’ll see you all at the British Embassy at 7:30.

  • [Meeting recessed]