This article provides an overview of the three episodes of quantitative easing (QE) pursued by the Bank of Japan (BOJ) since 2001. It begins with a brief account of the initial reluctant shift to unorthodox policies under BOJ Governors Hayami and Fukui in 2001-06 (here designated QE1) and then covers the equally reluctant adoption of QE by Governor Shirakawa in 2010-13 (QE2). The article then turns to an account of the attempt since April 2013 by the BOJ under Governor Kuroda, designated "quantitative and qualitative easing" (QQE), to revive the economy and achieve a 2 percent inflation target. None of these attempts at QE has been successful in raising the broad money growth rate for M2 sustainably above the 2-3 percent per annum range where it has languished for the past 25 years. Consequently, Japan's attempts at QE have all failed to raise the equilibrium level of Japanese nominal GDP by any material magnitude, and so far, attainment of the 2 percent inflation target under QQE has remained elusive. At the time of writing (October 2016), the Japanese economy therefore continues to grow at a low rate with periodic lapses into deflation. After discussing the case of Japan, the article compares the experience of the United States in 1929-33, when there was no QE, and the experience of 2008--14, when the Fed conducted QE over three periods. The comparison is deliberately focused on the quantitative aspects of the policy, not its interest rate effects. Finally, the article explains that there are two brands of QE, and that the failure of QE in Japan is fundamentally due to the choice of the wrong brand of QE. Given the type of QE that the Japanese authorities have chosen, the policy cannot be expected to succeed, except under limited conditions. (1) If QE were to be implemented according to a different design, the prospects of success would be much greater. In brief, the primary reason for the failure of BOJ-style QE or QQE derives from the habitual tendency to buy securities from banks instead of from nonbank private-sector entities (such as nonbank financial firms, nonfinancial firms, households, or foreigners). While QE policy in Japan boosts the monetary base, it does not increase broad money. But it is broad money that drives nominal GDP, not the monetary base.
BOJ's QE1: 2001-06 under Governors Hayami and Fukui
The Japanese economy experienced a classic asset bubble during the period 198.5-90, featuring steep rises in the prices of equities, real estate, and other assets such as golf club memberships. (2) The stock market peaked in December 1989 and the real estate market peaked in the July--September quarter of 1991. Those peaks were followed by steep declines in asset prices, culminating in contractions of real GDP in several quarters of 1992 and 1993 and deflation as measured by the CPI from July 1994.
During the 1990s, numerous types of countermeasures were adopted by the Japanese authorities to combat the economic downturn, such as a series of fiscal expansion plans (see Wright 2002), a half-hearted attempt to recapitalize the banks and sell oil the toxic assets from 1998, (3) cutting the BOJ's overnight policy rate to zero (ZIRP, or zero interest rate policy) by 1999, and intermittently allowing the Japanese yen to depreciate in an effort to promote export-led growth. All of these policies proved ineffective. The reason was that by September 1992, the money supply (M2) had declined on a year-on-year basis--unprecedented in postwar Japan--and it continued to grow only at a snail's pace, averaging just 2.5 percent per annum ever since 1992. (4) Without adequate growth of the broad money supply, nominal GDP remained in a prolonged slump, reflected in weak real GDP growth combined with persistent deflation.
After the BOJ adopted ZIRP in the spring of 1999, and after a second round of capital increases for the banks in March 1999, the economy started to perform a little better, with real GDP strengthening into 1999 Q4. Inflation remained negative but moved back toward zero. Influenced by these more favorable developments and by the arguments of crusading BOJ Policy Board member Eiko Shinotsuka that Japanese savers needed higher interest rates from the central bank, the BOJ decided to end ZIRP and increased the (targeted) uncollateralized overnight interest rate to 0.25 percent in August 2000. The rate hike was maintained until March 2001, but by then, stock prices had fallen back and a renewed economic downturn, was evident, accompanied by a recession in the United States following the bursting of the tech bubble. Faced with this further downturn, the BOJ lowered the overnight rate again and finally turned to a radical new proposal: quantitative easing.
The Policy Board statement of March 19, 2001, emphasized the extraordinary nature of its decision: "The Bank has come to a conclusion that the economic conditions warrant monetary easing as drastic as is unlikely to he taken under ordinary circumstances" (Bank of Japan 2001). In this first episode of QE, which started in March 2001 and ended in March 2006, the BOJ purchased a net 37 trillion yen of securities, expanding its balance sheet from 115.3 trillion yen to 152.3 trillion yen. Its purchases consisted of Japanese government bonds (JGBs) and short-dated financing bills or promissory notes (known as legal a).
However, as Figure 1 demonstrates, the increase in Japan's monetary base was not matched by any significant change in trend of the broad money supply, M2. Between 2001 and 2006, the monetary base expanded by 70 percent, but there was virtually no change in the trajectory of broad money (M2), while bank lending declined by close to 10 percent over the period 2001-06.
Even so, the policy started to have some success when stock prices increased by 57 percent between May 2005 and April 2006, although it should also be noted that there was a significant rally on Wall Street at the same time. The economy, too, had started to perform better, recording steady increases in real GDP during the six quarters from 2005 Q1 until 2006 Q2. Moreover, inflation returned to relative price stability. It can therefore be argued that QE1 was terminated prematurely.
Throughout the period 2001-06, the BOJ Policy Board members, with some exceptions such as the forthright Nobuyuki Nakahara, were far too timid, frequently expressing the desire to return to orthodox policies--by which they meant implementing monetary policy by adjusting short-term interest rates. Fundamentally, they failed to recognize that major balance sheet repair was needed across key sectors--households, and financial and nonfinancial corporations--and they therefore expected the economy to return to normality after only a brief interlude of unorthodox policy (Koo 2003).
In 2003-05, the BOJ had set three conditions for exiting QE, and by late 2005, these conditions were being met. Therefore, on March 9, 2006, the Policy Board decided to terminate QE, deciding to return to ZIRP while the outstanding excess reserves were reduced.
The speed with which the BOJ's balance sheet declined from April 2006 and the minimal impact this had on financial markets or the economy highlight a second aspect of the problem with the BOJ's brand of QE. Instead of buying only long-dated securities such as JGBs that would remain on the BOJ's balance sheet for an extended period, the BOJ purchased large amounts of tegata (short-term financing bills with maturities of less than one year), also primarily purchased from banks rather than nonbanks. Consequently, without the BOJ overtly selling any securities into the market, the maturing of these securities reduced the BOJ's balance sheet abruptly from 152 trillion yen in March 2006 to just 114 trillion yen, or by 38 trillion yen by the end of June 2006.
If BOJ-style QE purchases were supposed to have an expansionary effect on asset markets and the economy, a reduction in the BOJ's balance sheet should have led to seriously adverse or contractionary effects on the financial markets and the economy. Yet the drastic decline in the monetary base (or its counterpart, BOJ assets) between April and June 2006, unwinding the entire five-year buildup of QE within three months, had remarkably little impact on either the Japanese stock market or on the Japanese economy. Measured by the Nikkei index, equities continued to rise to a peak of 17,563 on April 7, 2006, and then declined to 14,751 by June 9, a decline of 16 percent. However, by February 23, 2007, almost a year after the end of QE1, the index had reached a new interim high of 18,188, or arise of 23 percent. Similarly, following the termination of QE1 the economy continued to grow until 2008 Ql, with only two negative quarters of real GDP growth before the 2008 global recession--in 2006 Q3 (followed by a strong upturn in Q4) and 2007 Q3. Inflation, too...