US monetary policy: QE3

Read the article in the Cayman Financial Review Magazine 

When Lehman Brothers collapsed on 15 September, 2008, the US interbank market through which banks with excess short-term liquidity lend to those banks in need of it stopped functioning.

Banks didn’t know which banks they could trust and chose to sit on their excess reserves instead. The Federal Reserve Banks stepped into the breach and dramatically increased their lending to the banks that had been cut off in the market. This is the liquidity support to the system that central banks are expected to provide and it saved American financial markets from almost certain melt down.

The Fed’s record since then has been more mixed. It introduced a number of non-traditional lending facilities and lent large amounts to non-traditional borrowers. It also undertook massive open market purchases of non-traditional assets, primarily Mortgage Backed Securities guaranteed by the Government Sponsored Enterprises (GSEs) and the debt of GSEs and then in 2010-11 longer-term Treasury securities.

The FOMC by late 2008 also lowered short-term interbank market interest rates, the Fed funds rate, from 5 per cent to almost zero (0.0 – 0.25 per cent), as shown in the Figure 1.


Given the international nature of the liquidity market for US dollars, the Fed also entered into currency swap agreements with major central banks under which the Fed sold as much as $600 billion US dollars for foreign currencies under repurchase agreements. More recently it has undertaken large purchases of long-term Treasury bonds and sterilised their impact on liquidity with sales of short-term Treasury bills (Operation Twist).

During this period monetary policy had two primary objectives. The first was to unblock the flow of liquidity in the financial sector that had resulted from the enormous uncertainty about the value of mortgage-backed financial products and the resulting collapse of confidence in the soundness of the key bank players.

An important contributor to the Fed’s inability to supply the needed liquidity was its practice of conducting its daily open market operations (purchases and sales of government securities by the Federal Reserve Bank of New York on behalf of the whole system) only with its so-called “primary dealers”. There are currently only 16 primary dealers who gain access to privileged market information in exchange for their obligation to bid in the Fed’s daily operations. These primary dealers intermediate the Fed’s operations with the rest of the financial system. With the advent of electronic trading and accounting the rational for this system no longer exists.

The collapse of the interbank market meant that the liquidity provided by the Fed to the primary dealers was not being spread more widely through the financial system. Many of the special lending facilities established by the Fed in this period were to overcome this problem.

When added to the funds banks with excess liquidity were not lending, the injection of additional funds to the banks needing it pushed excess reserves (bank deposits at Federal Reserve Banks in excess of required levels) to historically high levels (see Figure 2). The Fed instituted interest payments on these excess reserves to encourage banks to continue holding them and to give it a new instrument for controlling short-term interbank interest rates when it eventually needed to increase them.


The second objective was to lower interest rates in order to encourage investment and aggregate demand and to reduce the negative impact of mortgage interest rate resets on adjustable rate mortgages with very low introductory rates (remember those?). Government interference (attempts at support) with mortgage refinancing, restructuring and default continues to befuddle stabilisation of that market, but the housing sector is gradually recovering none the less.

Counter-cyclical variations in monetary growth and interest rates fall within traditional monetary policy operations, though they are not without potentially negative side effects. Low interest rates reduce the return on saving. Pension funds must set aside larger amounts to cover pension commitments.

Prolonged periods of low rates distort the allocation of investments potentially creating new asset bubbles like the real estate bubble that caused the current crisis. In addition, essentially zero fed funds rates have prevented the revival in the interbank lending market perpetuating the need for more active supply by the Fed.

However, the Fed went beyond these measures to support the market prices of specific asset classes. The Fed’s large purchases of Mortgage Backed Securities (MBS) was motivated by the belief that the market was overly discounting potential losses on this assets, which reduced their value in the portfolios of banks and others that held them.

Putting a floor on their prices helped limit bank losses and potential bank failures. It also reduced somewhat the market’s uncertainly over the value of these assets. The Fed’s judgment on these prices seems to have been vindicated and it is enjoying a profit on its purchase of MBSs at discounted prices. But the operation crossed the line into what is traditionally a fiscal operation and potentially weakens the independence of the Fed to conduct monetary policy.

In fact, purchasing MBSs in order to support their value in bank portfolios was the original and more appropriate objective of the Troubled Asset Relief Program (TARP). When the Treasury decided that it was too difficult to value these assets and redirected TARP funds to the direct recapitalization of banks, the Fed stepped in and undertook this fiscal action.

These actions were undertaken in the heat of crisis, but looking back it is clear that these purchases should have been undertaken by the Treasury as originally planned rather than by the central bank.

The Fed’s operation twist (swapping long dated Treasuries for short dated ones) is akin to a Treasury debt management function and further threatens central bank independence. Moreover, the Fed’s desire to drive down longer dated interest rates is bad debt management policy. The Treasury should be lengthening the maturity of its debt to lock in for a longer period the very low rates on longer maturities.

By lengthening the average maturity of the public debt held by the Fed, any budgetary savings to the Treasury of delaying the impact on its borrowing costs of the increase in interest rates that will certainly come, will to that extent be offset by income losses by the Fed. In other words, the cost to the government (combining the balance sheets of the central bank and treasury) of its borrowing depends on what the treasury issues not who holds it.

A central bank’s core activity is to supply its currency in amounts and at rates that are consistent with preserving the value of that currency. The market’s expectations for the future value of that currency (inflation expectations) are formed on the basis of the rules set by the central bank for regulating the supply of its currency and its track record of adherence to those rules.

The Federal Reserve suffers from a dual mandate to maximise employment as well as stabilise prices.

Maximum sustainable employment is determined by real factors in the economy that are not influenced by monetary policy, so the Fed has properly interpreted its mandate as minimising departures of employment from its maximum sustainable level (cyclical unemployment) in the course of seeking long run price stability.

In a statement issued by the Fed following the January 2012 meeting of its Federal Open Market Committee (FOMC) it explained its policy objectives as follows: “The Committee judges that inflation at the rate of 2 per cent (as measured by the annual change in the price index for personal consumption expenditures, or PCE) is most consistent over the longer run with the Federal Reserve’s statutory mandate.”1

The Fed’s policy choices for achieving these goals rests fundamentally on its choice of how rapidly to increase the money supply over time. If money supply growth matches the growth in its demand (over longer periods the demand for money tends to growth with income – GDP), monetary policy is neutral, neither stimulating nor restraining aggregate demand. For many years the Fed’s day-to-day tool (its so called operating instrument) for controlling the money supply was the strong influence its purchases or sales of assets exerted on the overnight interbank rate (the so called Fed Funds rate).

It can push the Fed Funds rate lower (than prevailing market rates, which ultimately reflect market fundamentals and inflation expectations) by supplying more of its currency (open market purchases of securities – traditionally mostly treasury bills), thus causing the money supply to grow more rapidly. It can slow monetary growth by draining liquidity (open market sales of treasury bills) and pushing the Fed Funds rate higher (above rates determined by market fundamentals).

The experiences of the last four years illustrate the challenge and difficulty in setting and interpreting monetary policy anchored by monetary or inflation targets rather than exchange rates (as in the old gold standard). The US housing bubble stalled in 2006 and burst in mid 2007. Real output growth (real GDP) slowed in the last quarter of 2007. As a result, the non-observable equilibrium real rate of interest (the rate that would keep the economy fully employed) began to fall.

Without a downward adjustment in the fed funds rate, the prevailing rate would become increasingly above the equilibrium rate (ie, money policy would become increasingly tight). In August 2007 the Fed responded to slowing growth with a series of increasingly large cuts in the funds rate from 5.25 per cent to 2.0 per cent by May 2008. This reversed the slowdown in M2 growth that had started in late 2007. M2 growth fell from 6.7 per cent in September 2007 to 5.7 per cent in January 2008 and increased back to 7 per cent in March 2008. Real GDP growth, which had turned negative the first quarter of 2008 (-1.8 per cent) revived to 1.3 per cent at a compounded annual rate in the second quarter. This should have been the happy start of a normal recovery.

However, the funds rate was kept at 2 per cent after May for four months (see Figure 3) and the Fed’s minutes and the public comments of some governors increasingly expressed concern over the inflationary potential of further funds rate cuts. As a result, market expectations of the future stance of monetary policy shifted toward tightening (an expected funds rate above the equilibrium real rate and thus a fall of the money supply below its demand). 


Real GDP growth reversed in the third quarter of 2008 registering a 3.7 per cent decline and a staggering minus 8.9 per cent in the fourth quarter at compounded annual rates following the collapse of Lehman Brothers on 15 September, 2008. M2 growth slowed to below 5 per cent at an annual rate and the depreciation of the dollar against the euro, which had declined almost 27 per cent from the beginning of 2006 to April 2008, was reversed by a 20 per cent appreciation between May and November (see Figure 4). The US had entered the Great Recession.2


Robert Mundell, Nobel prize laureate in economics and often called the father of the Euro, characterised the period as a dramatic tightening of policy and one of the greatest mistakes of the Fed’s history. In fact, he points to the wide exchange rate swings between the dollar and Euro of almost 100 per cent over the last decade as a serious problem for the global trading economy (see Figure 4). Imagine the difficulties posed by these swings for third countries, like the Cayman Islands, that trade with both the US and Europe.

The Fed responded to these developments and the collapse of Lehman Brothers in September 2008 by resuming its cuts in its fed funds rate target until it reached zero by the end of the year and doubling its balance sheet via large asset purchases. The economic response was a recovery in monetary growth (increasing to very high rates by the beginning of 2009), a retreat of the exchange rate’s overvaluation to more appropriate levels, and an end to the contraction in output (real GDP growth became positive in early 2009). However, with a zero funds rate target, it seemed that the fed had exhausted its ability to lower real market rates to or below a still depressed equilibrium real rate.

If it thought that further monetary stimulus was still needed to achieve its 2 per cent inflation objective (ie, to avoid deflation), the Fed could have continued open market purchase of financial assets even with the fed funds rate at zero (what has since come to be called Quantitative Easing – QE).

In fact, the Fed’s total assets did not grow further after the end of 2008 for over a year. The growth rate of M2 peaked in the beginning of 2009 gradually slowing to a more “modest” 8 per cent year on year rate by mid 2009 before plunging to almost zero by the end of 2009.

The impact of the money supply on economic growth and on prices reflects both money supply and money demand factors. If the response of GDP and inflation are “caused” by increases or decreases in the growth in the money supply, the responses are generally seen within 6 to 18 months (a one year lag on average, see Figures 3 and 5). 


In the case of the late 2009 fall in M2’s growth rate some of the drop in GDP growth occurred at the same time (real GDP growth of 1.4 per cent and 4.0 per cent in Q3 and Q4 of 2009 dropped to 2.3 per cent in the first quarter of 2010 and stayed at about that rate for the rest of the year), suggesting that it was not the result of the contemporaneous collapse of money growth.

However, the drop in the real growth rate to zero in the first quarter of 2011 is consistent with its being “caused” by the one year earlier collapse in M2 growth. In other words the evidence of monetary policy’s impact on income is mixed.

As seen in Figure 6 showing the level of M2, the sharp drop in its growth rate in the second half of 2009 reflects a temporary pause after its very rapid growth in 2008 followed by resumed growth. It is not obvious that such behaviour should have much effect on real GDP. 


More generally the attribution of the behaviour of M2 to the subsequent behaviour of real output and inflation has become less reliable since the 1970 because of difficult to forecast shifts in the demand for money of specific types as the result of financial liberalisation and technological advances in payment technologies.

Steve Hanke, Professor of Applied Economics at The Johns Hopkins University in Baltimore and a strong advocate of dollarization or currency boards, argues that money growth has been too slow since the financial crisis of 2007 up to the present and that monetary policy has been a drag on the economy throughout and thus contributed to our slow recovery3 (when people hold less money than they find optimal – not to be confused with income – they must consume less of other things in order to build up their money balances).

While he focuses on the St Louis Fed’s Monetary Services Index 2 (also known as the divisia index for M2 – see my review of William Barnett’s “Getting it Wrong” in this issue of the Cayman Financial Review), its behaviour has been almost identical to the regular M2 series in recent years. Over the past year ending in September of 2012, however, M2 grew modestly above its long run average of 5.5 per cent per annum.

If M2 growth has been slower than the growth in its demand, ie if money is too tight and slowing growth, interest rates would be above their normal or long run equilibrium levels. Though interest rates in the US and Europe are at historic lows (see Figure 7), suggesting that money is not tight, the equilibrium real rate cannot be directly observed. 


If credit is constrained it is more likely to be for structural reasons (discussed above) or because of market uncertainty and the resulting high risk premiums that price borrowers out of the market (see below).

The Fed’s Quantitative Easing policies are simply a new name for an old and traditional policy tool. When the Fed cannot expand the supply of its currency further by setting its interest rate (federal funds rate) target lower, because it is already at zero, it continues buying securities in the market at the existing zero funds rate.

This is QE 1, 2 and 3. The Fed manipulates the supply of its monetary liabilities by injecting reserves when it purchases securities (“printing money” or more generally called open market operations) or withdrawing (draining) it when it sells securities back to the market.

QE3 resumes the purchase of Mortgage Backed Securities “and other securities” but this time for an indefinite period – for “as long as it takes to reduce unemployment to acceptable levels”. Its goal is to further lower already very low mortgage rates and thus increase housing demand and housing prices. It is also expected to increase demand for other financial market assets (to replace the ones purchased by the Fed as part of QE3) raising stock prices and wealth more generally.

The increased housing demand should also directly promote increased housing construction. All of these factors, it is hoped, will increase household spending and thus aggregate demand leading to increased employment. Empirical studies are increasingly indicating that the stimulus from the successive QEs are small and temporary.4

In my opinion this policy is a mistake. The economy will recover only when households and firms pay down their excessive debt (deleverage) to more comfortable and prudent levels and when they know the tax and regulatory rules under which they can invest. Increases in exports as the result of a more depreciated (less overvalued) dollar, have helped make up for some of the lack of household and firm demand for output, but current economic weakness in Europe is now hurting our exports. There is also considerable risk that any impact of the Fed’s stimulus will come when it is not needed (a year or so from now) and fuel the next round of excess.

The government’s efforts to prop up demand with government stimulus and the Fed’s low interest rates slowed the adjustment that is needed before recovery is possible. Government interference with the normal market handling of underwater and/or non performing mortgages continues to slow the needed adjustment to housing prices, though the bottom has finally been reached and that market, both new construction and existing houses, is gradually recovering.

The biggest remaining impediment to increased growth and employment is the market’s uncertainty over the details of the resolution of the fiscal cliff toward which the federal budget is heading (and other government policies effecting the profitability of investing) in an environment of political dysfunction.

Estonia and Lithuania provide dramatic examples of countries that adjusted quickly and recovered quickly. Lithuania has a currency board whose currency is fixed to the euro and Estonia was a currency board fixed to the euro but for the past year has been part of the euro zone. Thus both countries have money supplies determined automatically by market demand, similar to the way the gold standard worked. Both undertook immediate and dramatic corrective actions (austerity) to address their financial problems arising from the 2008 global Great Recession, Lithuania as part of an IMF supported program and Estonia on its own.

Both suffered a large drop in real GDP in 2009 (-14.3 per cent in Estonia and -14.8 per cent in Lithuania). Both are now recovering rapidly (in 2010 Estonia grew at a 2.3 per cent rate and Lithuania at 1.4 per cent and in 2011 Estonia grew at 7.5 per cent and Lithuania at 6.3 per cent). While real growth slowed for both in 2012 it remains well above the European average (Estonia 2.4 per cent and Lithuania 2.7 per cent compared with Germany at 0.9 per cent).

The Federal Reserve is now pushing on a string that it can do very little to affect demand. On the contrary, inflation expectations increased about 25 points upon its announcement of QE3 and its continued open market purchases only increase the difficulty of unwinding its huge holding of securities and the banking system’s huge excess reserves without becoming inflationary.5

While the Fed has the tools needed to reduce its huge holdings of securities and to drain banks’ large excess reserves, moving its policy interest rates above the increases in market equilibrium interest rates that recovery and the ballooning public sector debt will bring about, will be politically difficult.

The recovery of the housing sector, the sector that caused the Great Recession in the first place, is already picking up. QE3, if it has any effect at all, risks over fuelling the housing recovery and reintroducing new distortions and bubbles. One important thing that is needed, along with the natural corrective forces of time, is what has been needed for many years, a credible set of debt reduction measures that the market can count on now to gradually rain in and reverse fiscal deficits and ultimately the relative size of the outstanding and prospective debt over the next decade.

This will consist largely of reducing future entitlement outlays (Medicaid, Medicare, social security), which are set to grow dramatically with an aging population, reducing the world’s largest by a very large margin military budget and enacting a more efficient and equitable tax system.

Everything must be on the table. Our representatives must move beyond their election battles and agree on a set of clear, sensible, and predictable policies that will eliminate deficits in the future and contain the debt/GDP and provide the basis for businesses to invest and households to spend within their means. I have written about all of these before.6

The economy needs much greater certainty about government policies and the rules of the game before investment will resume at sufficient levels. On top of the US government’s inability so far to convincingly address its long run (but no longer future) demographically driving debt problems (the baby boomer problem has only be made worse by recklessly expanding government programmes and commitments), and the proliferation of costly and ever changing regulations, uncertainty over the central bank’s policies and the future value of the dollar domestically and internationally undermine almost every aspect of the economy.

The Federal Reserve has become admirably transparent. Earlier efforts to provide the market with better guidance by indicating the FOMC’s intention to keep the funds rate near zero for at least another two years only signalled the market that the Fed thought the economy would not recover until then. The latest innovation in the Fed’s market guidance is a significant improvement and is much welcomed.

In the 12 December statement of the FOMC’s latest decision the Fed reported that “the Committee decided to keep the target range for the federal funds rate at 0 to 1/4 per cent and currently anticipates that this exceptionally low range for the federal funds rate will be appropriate at least as long as the unemployment rate remains above 6-1/2 per cent, inflation between one and two years ahead is projected to be no more than a half percentage point above the Committee’s 2 per cent longer-run goal, and longer-term inflation expectations continue to be well anchored.”

Policy decisions affect the economy in the future, generally one to two years out. Thus policy must be made on the basis of model forecasts. The above announcement shares with the public the decision criteria imbedded in such model forecasts. This transparency is helpful and very welcomed.

However, as the above discussion of the past four years should make clear, even the most independent and skilled central bank with floating exchange rates and a price stability anchor has an almost impossible task. Even rule based monetary control tends to aggravate rather than dampen asset price bubbles and cycles in economic activity. It is time to give up the experiment and return to a real anchor to monetary policy.

The gold standard was such a system. It had weakness, which can be overcome. The world needs to return to currencies whose value is fixed to some measure of the real economy such as described in my Real SDR Currency Board.7


  2. Scott Sumner, “The real problem was nominal”, Cato Institute, September 14, 2009. 
  3. Steve Hanke, “It’s the Money Supply Stupid”, Globe Asia, July 2012.
  4. See Jonathan H. Wright, “What does Monetary Policy do to Long-Term Interest Rates at the Zero Lower Bound?” May 9, 2012,
  5. Larry White, “The real problem today is not so much nominal”, 23 September 2012
  7. Warren Coats “Real SDR Currency Board”, Central Banking Journal XXII.2 (2011), and “Do We Need A New Global Currency?”, Cayman Financial Review, Issue 18 First Quarter, 2010.