May 29, 2009
A new accord?
As days go in the U.S. Treasury market, Wednesday was a rough one. Bloomberg News described the day's developments as follows:
"The difference in yields between Treasury two- and 10-year notes widened to a record on concern surging sales of U.S. debt will overwhelm the Federal Reserve's efforts to keep borrowing costs low. …
"The unprecedented government borrowing has created concern about a rise in consumer prices. Policy makers have expanded the Fed's balance sheet to $2.2 trillion while excess reserves at U.S. banks have increased to $896.3 billion.
" 'Inflation is in the headlight of many investors,' wrote Andrew Brenner, co-head of structured products and emerging markets in New York at MF Global Inc., in a note to clients today."
By cosmic coincidence, I was reading that story in Tokyo as I prepared to chair a session at the Bank of Japan's 2009 International Conference on Financial System and Monetary Policy Implementation in which Marvin Goodfriend (Carnegie-Mellon professor and former Richmond Fed policy adviser) presented his thinking on keeping the central bank's inflation objectives firmly in hand at a time of rapidly rising government debt and large increases in the Fed's balance sheet. Professor Goodfriend's case is laid out in a paper titled "Central Banking in the Credit Turmoil: An Assessment of Federal Reserve Practice" (which was also presented at a conference devoted to research on the interactions between monetary and fiscal policy—that is, government spending and tax—co-sponsored by Princeton University's Center for Economic Policy Studies and Indiana University's Center for Applied Economics and Policy Research). Goodfriend pulls no punches:
"The 1951 'Accord' between the United States Treasury and the Federal Reserve was one of the most dramatic events in U.S. financial history. The Accord ended an arrangement dating from World War II in which the Fed agreed to use its monetary policy powers to keep interest rates low to help finance the war effort. The Truman Treasury urged that the agreement be extended to keep interest rates low in order to hold down the cost of the huge Federal government debt accumulated during the war. Fed officials argued that keeping interest rates low would require inflationary money growth that would destabilize the economy and ultimately fail.
"The so-called Accord was only one paragraph, but it famously reasserted the principle of Fed independence so that monetary policy might serve exclusively to stabilize inflation and the macroeconomic activity. …
"The enormous expansion of Fed lending today—in scale, in reach beyond depository institutions, and in acceptable collateral—demands an accord for Fed credit policy to supplement the accord on monetary policy. A credit accord should set guidelines for Fed credit policy so that pressure to misuse Fed credit policy for fiscal purposes does not undermine the Fed's independence and impair the central bank's power to stabilize financial markets, inflation, and macroeconomic activity."
Goodfriend's "new accord" amounts to asserting a set of principles that would reinforce price stability as the central goal of monetary policy, set the Fed down the road of extricating itself from the extraordinary credit market interventions of the past year-and-a-half, and join the central bank and Treasury in common cause toward the goal of doing everything possible to make sure that the Fed and Treasury don't go there again.
The paper was provocative, but it also raised a couple of fundamental questions. In particular, what is the degree of autonomous fiscal risk-taking appropriate to allocate to a central bank? If such powers are granted, how often and under what conditions ought those powers be exercised? And how far should the powers reach? If, as Goodfriend states, "the central bank's power to stabilize financial markets, inflation, and macroeconomic activity" requires lender-of-last-resort interventions—and hence pure credit policy interventions—what does that imply about the appropriate scope of monetary and regulatory authorities going forward? If the "shadow banking system" is the de facto banking system of the modern era—as Yale University's Gary Gorton argued in a paper presented at the Atlanta Fed's annual Financial Markets Conference held earlier this month—is the central bank's lender-of-last-resort function meaningful if narrowly construed (that is, if it fails to reach the shadow banking system)?
These, really, are first-order questions. On to the debate.
By David Altig, senior vice president and research director of the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference A new accord?:
May 28, 2009
Housing starts, remittances and macroeconomic developments
Recent evidence collected by the Dallas Fed's Pia Orrenius suggests that apprehensions of undocumented workers attempting to cross the U.S.–Mexican border are a good predictor of the overall American job market. Simply put, if one wanted to predict job market conditions in July of a given year, one should examine immigrant apprehensions in January. Orrenius finds that more immigrants attempt to cross the border from Mexico (and more of them are caught doing so) when immigrants believe the U.S. economy would offer more jobs in the near future.
One area of the economy that relied heavily on immigrant labor was housing. The following chart plots monthly U.S. housing starts (lagged five months) and remittances to Mexico. (I use year-over-year growth rates and smooth the noise from very short-run fluctuations by using a three-month moving average in my analysis.) I use remittances as a proxy for migrant Mexican labor.
Figure: Housing Starts and Remittances to Mexico
Note: Remittances in U.S. dollars. Housing starts indicate new, privately owned housing units.
Source: Bank of Mexico (remittances), Haver Analytics (housing starts)
The correlation between the two data series is strikingly high. For instance, the plunge in housing starts that began in early 2006 was followed by a sizable drop in remittances growth five months later. Of course, the results are not unexpected as the construction industry heavily employs immigrant labor.
Also, it is well known that immigrants send remittances to their country of origin on a regular basis. Some estimates indicate that the remittances sent by immigrants from developing economies back home reached $305 billion in 2008. (As an aside, keep in mind that because of unrecorded immigration flows through formal and informal channels, the actual numbers are likely to be significantly larger.) Remittances are particularly important for smaller Latin American countries. In 2007, for instance, recorded remittances represented more than 10 percent of the gross domestic product in several Latin American countries, including Honduras (25 percent), Guyana (24 percent), El Salvador (16 percent), and the Dominican Republic (13 percent), among others.
What is especially remarkable from a macroeconomic perspective is the volatility of these capital flows. During the housing boom, remittances to Mexico were growing at 20–25 percent annual rates (see the chart). With the onset of the global economic crisis, however, remittances have been declining, falling by almost 10 percent early this year. For smaller emerging economies, the volatility in remittance flows becomes a significant extra source of instability.
Migrant workers enter the country in response to upturns in domestic labor demand, and that upturn results in higher remittances both because of the increased number of immigrants but also because the existing stock of immigrant workers is earning more. Conversely, a downturn in labor demand should be reflected in lower remittance flows because of out-migration as workers return home and because of lower earnings among the remaining stock of migrants. But what happens when some of those workers have entered the country illegally?
In a study published in 1997, Belinda Reyes found that about two-thirds of the undocumented immigrants returned to Mexico within three years upon arrival. In a recent paper I wrote with Andrei Zlate, we explore the implications of changes to enforcement policies for the U.S./Mexican border on undocumented labor and remittances. We find that increased border enforcement during the last decade has broken the typical pattern of flows of undocumented workers. Basically, while increased enforcement makes it harder/more expensive to enter the country, it also reduces the incentive for those already in the country to leave. Why? Because of the high cost/risk associated with reentering the United States in the future.
In a recent paper, Carolina Rodriguez-Zamora adds support to our claims. She finds that as the U.S. Department of Homeland Security increases the amount of resources spent policing the border undocumented immigrants tend to stay longer.
Increased enforcement protects the existing stock of undocumented immigrants from additional competition, and this development can put upward pressure on wages when U.S. labor demand is high. When labor demand is low, rather than returning home, these individuals could remain in local labor markets, placing additional downward pressure on wages.
By Federico Mandelman, research economist and assistant policy adviser at the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference Housing starts, remittances and macroeconomic developments:
May 21, 2009
More on interest on reserves
Following my previous macroblog post on tools for managing the Federal Reserve’s balance sheet, I received a few questions, and I’m using today’s post to reply to these.
First, a question from Alexander Singer:
Wouldn't the reserves we are talking about fall as well as a result of banks actually lending money out to their customers?
The banking system cannot create or destroy reserves no matter how many loans they make as long as the borrowed funds are deposited. For example, if one bank takes a portion of its reserve balance and turns it into a loan, and those borrowed funds are deposited at a second bank, then that deposit is matched by an equivalent increase in the second bank’s reserve balance. The net effect on reserves is zero.
The Fed has created a substantial amount of reserves during the past year and a half. At the end of 2007 reserve balances at Federal Reserve stood at about $8 billion. Currently reserve balances are closer to $900 billion.
Will the broader money supply grow once lending increases? Yes, lending will generate bank deposits, and bank deposits (plus currency) equals money. But that won’t have any direct impact on total reserve balances within the banking system.
A second question comes from Tom:
With interest on reserves (IOR), why would a bank not want to keep its reserves at the existing level after the policy rate is increased? They are operating on a horizontal segment of their demand for reserves. I just don't see why raising the policy rate will be a problem with IOR. What am I missing?
Some economists have argued that paying interest on reserves will render the demand for reserves indeterminate (see, for example, the 1985 JME article by Neil Wallace and Tom Sargent). But proponents argue that pinning down the demand function is not crucial for an interest on reserves based monetary policy to be effective (see, for example, Goodfriend 2002). In Goodfriend’s view (which is based on the assumption that interest rate rules for monetary policy deliver coherent outcomes for inflation and output), there is great value in a central bank being able to pursue separate interest rate and bank reserve polices. Interest rate policy would be used to maintain overall macroeconomic stability, while bank reserve policy would be used to address financial market objectives. Linking the policy rate to the interest rate paid on reserve balances means that a change in the interest rate does not require changing the supply of reserves.
Why might this be useful? Here’s one hypothetical scenario: Suppose the Fed needed to keep bank reserves at a high level because of lingering demand for liquidity in financial markets that is not being provided by the private sector. Now, suppose the Fed also wanted to tighten monetary policy because of separate macroeconomic stability concerns. Interest on reserves provides a tool to meet both a financial stability objective (by helping the functioning of credit markets) and a macroeconomic stability objective (by influencing banks’ willingness to lend to private borrowers).
A bit more detail. Banks used to view reserves as a “hot potato.” Reserves are useful to banks in making settlements, etc., but banks did not want hold too many reserves because they were a nonearning asset. The Fed did not compensate banks for holding the reserves and so banks had better uses for their funds. The Fed was able to keep the market price for reserves (its policy instrument) positive by keeping the amount of reserves scarce. But today, reserves are far from being scarce, and the Fed keeps the market price for reserves positive by paying interest on reserve balances (albeit only 25 basis points).
Keister, Martin, and McAndrews (2008) called this new approach “Divorcing Money from Monetary Policy.” I think that is a bit strong, but it does amount to an amicable separation between reserves management and monetary policy. Textbook descriptions of money and banking emphasize an important role for reserve management in the implementation of monetary policy (daily open market operations, etc.). But clearly that is not the only possible operating approach.
By John Robertson, vice president in Research at the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference More on interest on reserves:
May 14, 2009
Tools for managing the Fed's balance sheet
In his question and answer session following a speech to the Chicago Fed's Conference on Bank Structure and Competition on May 7, one of the topics Federal Reserve Chairman Ben Bernanke addressed was various tools for managing the Fed's balance sheet and the implementation of monetary policy in the future.
As I see it, the issue is that the Fed has created a very large amount of reserves and its conventional monetary policy tool, the federal funds rate target (the desired market rate for bank reserves), is very close to zero. If the funds rate target is going to be raised at some point in the future then either the Fed needs to be able to simultaneously keep demand for reserves at a high level or it has to reduce the amount of reserves available in order to drive the market rate higher. In other words, a positive interest rate would require eliminating a large portion of the excess reserves or elevating the demand for reserves in a way that is consistent with the market rate near the desired (target) level.
The options mentioned by the chairman last week fall along these two lines. One possibility is to have the demand for excess reserves at a high level, with the interest rate paid on reserves being the primary lever for the implementation of monetary policy. Last week I wrote about interest on reserves in the context of New Zealand's floor system and the fact that a large level of bank reserves replaced central bank daylight overdraft facilities for the purposes of meeting payment system needs. I also noted that U.S. banks have been putting their large holdings of reserves to similar use—significantly replacing daylight overdrafts. Nonetheless, it remains an open question as to whether market rates could be controlled satisfactorily by adjusting the interest rate paid on excess reserves alone given the size and complexity of the U.S. banking system.
What about reducing reserves? Beyond the decline in reserves that would automatically accompany reduced demand for the various Fed liquidity facilities, one tool for reducing reserves is outright sales of assets—Treasury securities, agency debt, and agency mortgage-backed securities—held in the Fed's portfolio. Doing so would reduce the level of bank reserves and also reduce the overall size of the Fed's balance sheet by the same amount.
Another option mentioned by the chairman is to use term reverse repos against these same Fed assets. A repo transaction resembles a collateralized loan, but it is technically a purchase and subsequent sale agreement with the price differential effectively reflecting the interest on the transaction. A Fed repo operation temporarily buys acceptable securities in exchange for reserves, while a reverse repo temporarily sells Fed-owned securities in exchange for reserves. To the Fed, a repo is an asset on its balance sheet that is matched by an increase in reserves on the liabilities side. In contrast, a reverse repo is a Fed liability that is matched by a decrease in reserves. Hence, a reverse repo does not reduce the size of the balance sheet, but it does reduce the amount of reserves. (Central bank trivia: Unlike most other central banks, the Fed convention is to talk about repos and reverse repos from the perspective of the counterparty.)
One option not mentioned by the chairman last week but talked about by others (see, for example, Janet Yellen's speech on May 6) is for the Fed to seek authority to issue its own unsecured debt. These Fed bills would be similar in effect to a reverse repo—replacing reserves with interest-bearing Fed obligations, but these obligations would not be backed by Fed-owned collateral. Many other central banks have the authority to issue their own unsecured short-term debt, and a number of central banks utilized this authority to manage their balance sheet during the heat of the financial crisis in 2008. For example, sterling bills and reserve bank bills have been used in the United Kingdom and New Zealand, respectively, as the primary means of draining reserves when reserves were created as a result of short-term liquidity facilities.
However, it is important to note that the central banks that introduced bills were generally facing large increases in reserves and declining amounts of government debt on their balance sheets. This situation significantly limited their ability to conduct large reverse repo operations or outright sales of securities to drain excess liquidity. For example, the Bank of England's holdings of government debt at the beginning of 2009 were less than 2 percent of its balance sheet, down from about 11 percent in early 2008.
In principle, then, it seems the Fed has plenty of tools available for managing reserves and the federal funds rate in the future. As my former countrymen in New Zealand are apt to say—no worries, mate! Of course, that leaves plenty of other things to worry about.
By John Robertson, vice president in Research at the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference Tools for managing the Fed's balance sheet:
May 06, 2009
When is rapid growth in a central bank's balance sheet not cause for concern?
The unprecedented growth in the Federal Reserve's balance sheet during the past year has generated considerable debate about potential problems for the economy down the road (see for example, here, here, and here). But a big change in the size of a central bank's balance sheet in a relatively short space of time is not necessarily a precondition for problems down the road. A case in point is New Zealand circa 2006.
In July 2006, the Reserve Bank of New Zealand (RBNZ) changed the monetary policy operating system from a channel or corridor system (like that used in Australia and Canada) to a floor system (see Neild 2006 for a description of this transition). Under this floor system, the RBNZ stopped offering free collateralized daylight credit to banks for settlement purposes. In other words, they removed the distinction between daylight and overnight reserves. Also under this new system, reserves were remunerated at the official cash rate (OCR), the RBNZ's target interest rate. Banks have access to RBNZ credit if needed as well, but at a rate 50 basis points above the OCR.
By the end of 2006 the target supply of bank reserves had increased sufficiently to allow for the smooth operation of the New Zealand payment system. The new level fluctuated around NZD 8 billion and represented an increase of 400 times the level under the previous regime. Todd Keister, Antonie Martin, and James McAndrews from the New York Fed have an interesting article describing the economics of a floor system. In that paper the authors stress that a floor system severs the link between the quantity of reserves and the target interest rate. A central bank could increase the supply of reserves—either for settlement or liquidity purposes—without changing the stance of monetary policy (the target interest rate).
Well, that's the theory. What about in practice? Did New Zealand's economy collapse under the weight of an inflationary spiral created by an explosion in the central bank money? In short, the answer is no. Because these newly created reserves were staying within the banking system there was no upward pressure on the broader money supply. For instance, M1 (currency plus checkable deposits) was NZD $22.9 billion in July 2006, and a year later it stood at $22.2 billion—a change that would not scare even the most hardcore monetarist.
Could a floor system work in the United States? Possibly. For one thing, with total reserve balances at the Fed about 18 times as large as they were a year ago there has been a sharp decline in demand for daylight overdrafts. Average daylight overdrafts for funds were $52 billion in the first quarter of 2008, but a year later that level had fallen to around $8.9 billion. With so many reserves in the system, the need for intraday borrowing from the Fed has decreased sharply. Of course, there are some big differences between the financial systems of New Zealand and the United States, including the fact that not all institutions depositing funds with the Fed are eligible to earn interest on reserves. But I do think the floor system provides some interesting food for thought—kiwifruit, perhaps.
By John Robertson, a vice president in the Research Department at the Atlanta Fed
TrackBack URL for this entry:
Listed below are links to blogs that reference When is rapid growth in a central bank's balance sheet not cause for concern?: