«I. Introduction The Institute for Monetary and Economic Studies (IMES) of the Bank of Japan (BOJ) held the 2014 BOJ-IMES Conference, entitled ...»
The discussant, Jan Marc Berk (De Nederlandsche Bank), argued that the ECB’s unconventional policy was not pure-QE and diﬀered from that by the Fed and pointed out that the author’s empirical setup and interpretation could be questioned. Regarding the unconventional policies introduced by the ECB, such as the long-term reﬁnancing operation (LTRO) and the Securities Markets Programme (SMP), he noted that these policies had addressed impairments in the monetary transmission mechanism in the euro area, which structurally diﬀered from that in the United States: the ﬁnancial intermediation system was market-based in the United States, while it was bank-based in the euro area. Regarding the interpretation of the author’s estimation results, he pointed out that the transmission mechanisms of QE, the exchange rate channel in particular, were not necessarily clear because there were numerous endogeneity issues and omitted factors in the empirical setup. Finally, Berk emphasized the importance of the distributional eﬀect of QE and suggested that it was crucial to address how to deal with implications of this eﬀect for the reputation, independence, and credibility of central banks.
From the ﬂoor, Obstfeld argued that the goals of the LTRO and the SMP were to support sovereign debt prices and strengthen bank balance sheets, so they were very diﬀerent from pure-QE. Yves Mersch (European Central Bank) commented that the situation in Europe today was quite diﬀerent from that in the United States during the recent ﬁnancial crisis, so the direct eﬀect of unconventional policy on asset prices would be quite diﬀerent. Paul Tucker (Harvard University) posed the question of why the portfolio balance channel would not work in the euro area given the presence of long-term investment institutions with a demand for longish-term bonds. Thomas J.
Jordan (Swiss National Bank) asked whether the impact of QE was linear or showed something like diminishing returns. Ito answered that it might be nonlinear and Sshaped, with an increase in the beginning but diminishing returns later. Kazuo Ueda (University of Tokyo) asked why the BOJ’s recent QQE had had diﬀerent eﬀects on Japan’s economy, asset prices in particular, from its QE during 2001–06. Ito answered 5 that the diﬀerences in the policy eﬀects were mainly due to the existence of the inﬂation target of 2 percent and the purchases of long-maturity bonds instead of short-maturity bonds. Kazumasa Iwata (Japan Center for Economic Research) asked whether the recent increase in Japan’s consumer price index was due to a rise in import prices stemming from the yen depreciation. Ito answered that a price increase had been observed for not only tradable goods but also other wide-ranging goods and services. R. Anton Braun (Federal Reserve Bank of Atlanta) suggested the use of the excess reserve ratio in the author’s empirical analysis.
B. Reserve Requirement Policy over the Business Cycle7 The recent ﬁnancial crisis has triggered an intense debate on the pros and cons of using macroprudential policy for macroeconomic stabilization purposes. Carlos A. Végh (Johns Hopkins University) focused on reserve requirement policy (RRP) as a macroeconomic stabilization tool and presented empirical results, in order to contribute to the discussion on macroprudential policy. The dataset used in the analysis covered 52 countries during 1970–2011. He ﬁrst reported the following stylized facts. About two-thirds of developing countries in the sample tended to change reserve requirements more than once in one business cycle, versus only one-third of industrial countries (and no industrial countries since 2004). Most developing countries used RRP countercyclically as a macroeconomic stabilization tool. The ﬁnding was based on the positive correlation between the cyclical components of real GDP growth and reserve requirements in these countries. The ﬁnding was also related to the fact that only slightly more than half of the developing countries had engaged in countercyclical monetary policy.
He argued that these countries were reluctant to use monetary policy countercyclically because of “fear of free falling” in their currency value in bad times and “fear of capital inﬂows” in good times. On the one hand, if interest rates were lowered in bad times, the currency could depreciate rapidly. On the other hand, if interest rates were raised in good times, too much capital could ﬂow into the country. Both rapid currency depreciation and massive capital inﬂow could ultimately destabilize the economy. Next, Végh reported the results on complementarity/substitutability between monetary policy and RRP. The most common policy mix for developing countries was acyclical interest rate policy and countercyclical RRP. He reported the estimation results on expanded Taylor rules that included a nominal exchange rate as well as inﬂation and an output gap and argued that RRP acted as a substitute for interest rate policy in these countries.
The estimation suggested that RRP was countercyclical while an interest rate did not respond systematically to output ﬂuctuations but tended to increase when the currency depreciated.
The discussant, Michael P. Leahy (Board of Governors of the Federal Reserve System), ﬁrst expressed appreciation for the dataset’s construction and the empirical analysis based on it. Then he discussed the authors’ ﬁndings on the use of RRP. First, regarding the ﬁnding on countries that employed RRP as a macroeconomic stabilization tool, it would be worth investigating the central banks’ motivation for changing reserve requirements by reviewing the banks’ announcement, documentation, and inFor details, see Federico, Vegh, and Vuletin (2014).
6 MONETARY AND ECONOMIC STUDIES / NOVEMBER 2014 Monetary Policy in a Post-Financial Crisis Era terviews. Second, regarding the ﬁnding on activist countries that used RRP in a countercyclical way, due attention should be paid to the long and variable lags of the RRP eﬀect in estimating the correlation between GDP and RRP. Focusing only on contemporaneous correlations might mislead the authors in their judgment of whether RRP was countercyclical. Next, he expressed some thoughts and conjectures stimulated by the paper. First, the paper could be interpreted as being about the choice of monetary policy instruments rather than about the use of a macroprudential tool. Second, RRP could not necessarily be separated from monetary policy. An increase in the reserve requirement puts upward pressure on interest rates and can have the same eﬀect as an increase in interest rates. Perhaps RRP might be more successful in some countries than interest rate policy because RRP acted as a countercyclical ﬁscal policy tool that imposed a tax on the banking sector, especially in countries with a poorly functioning ﬁnancial system. Finally, it was important to clarify what operation was used to implement interest rate policy as an alternative to RRP, because central banks could adopt diﬀerent kinds of operations to inﬂuence interest rates, including RRP.
From the ﬂoor, Mehmet Yörüko˘ lu (Central Bank of the Republic of Turkey) g mentioned the policy approach of his bank, where RRP was used to address credit growth and capital inﬂows and interest rate policy was used to deal with inﬂation. Tobias Adrian (Federal Reserve Bank of New York) asked whether RRP was eﬀective in stabilizing the economy and why RRP was used much less in industrial countries.
Obstfeld asked whether RRP would fade as emerging economies developed their capital markets and nonbank sources of lending. In reply, Végh referred to another paper of his that showed the eﬀectiveness of RRP in reducing credit growth and GDP in ﬁve Latin American countries. Regarding the inactive use of RRP in industrial countries, he mentioned that many industrial countries did not need to worry much about a free fall in exchange rates and that there was no point in using both interest rate policy and RRP if they were substitutes in these countries. Naoyuki Yoshino (Asian Development Bank Institute) pointed out that the transmission speed might diﬀer between RRP and interest rate policy because RRP inﬂuenced the banking sector immediately and capital markets gradually, while interest rate policy inﬂuenced immediately not only banks but also capital markets. Tucker asked about the importance of what seemed to be an implicit assumption that reserves were non-interest-bearing in the authors’ analysis, because central banks in advanced economies had in recent years moved to paying the policy rate of interest on reserves. Végh replied that his empirical results should hold qualitatively as long as interest on reserves was less than the short-term interest rate.
C. Banking, Liquidity and Bank Runs in an Inﬁnite Horizon Economy8 Nobuhiro Kiyotaki (Princeton University) presented a model that featured both ﬁnancial frictions and bank runs due to liquidity mismatch. He emphasized the complementary nature of a ﬁnancial accelerator mechanism and runs on banks, which was key to understanding the Great Recession of 2008–09 in the United States. In response to a negative shock to the economy, a ﬁnancial accelerator mechanism exacerbated the deterioration of banks’ balance-sheet conditions, which in turn raised the likelihood of
8. For details, see Gertler and Kiyotaki (2014).
7 bank runs. In particular, high bank leverage and a low liquidation value of bank assets were associated with a high likelihood of bank runs. Bank runs, if they occurred, caused a ﬁnancial crisis and further distress in the real economy. He mentioned that even if bank runs did not materialize, an increase in the likelihood of bank runs had negative eﬀects on the economy because it caused an increase in the cost of raising funds and a decrease in banks’ net worth, which were exacerbated by ﬁnancial frictions. He then discussed the role of ex ante and ex post policies to address ﬁnancial crises. In the model’s framework, bank runs could be eliminated by introducing deposit insurance. In practice, however, it was diﬃcult to introduce such insurance at all types of ﬁnancial institutions. In addition, trying to do so might induce the moral hazard of risk-taking. Capital requirements could reduce both risk-taking and the likelihood of bank runs, but could increase the cost of intermediation to the extent that raising equity capital was costly. In light of these considerations, he argued that addressing ﬁnancial crises required not only these ex ante policies but also ex post ones such as lender-of-last-resort functions and asset purchases by central banks.
The discussant, Braun, commented that the presented model could account for the chronology of the recent ﬁnancial crisis in the United States: a slowdown in economic activity, an increase in bank leverage, a rise in susceptibility to bank runs, and a sudden large and persistent decline in economic activity caused by bank runs. He then raised a question about who were the banks in the model in view of the recent ﬁnancial crisis and considered potential candidates. Money market funds and a company such as Bear Stearns were less likely to be the banks in the model, because their behavior diﬀered from the model’s assumptions and predictions. He argued that Lehman Brothers might correspond to the banks in the model, because the model captured an increase in the cost of raising funds and a sudden collapse in lending through repo markets in September 2008. Next, he commented on modeling issues. In particular, the model was so simple that it described only runs on all banks or no runs at all. He added that modifying the model to feature doubling down, as seen in the case of Bear Stearns, would enrich the behavior of banks in the model. Finally, he reﬂected on policy issues and pointed out a diﬀerence between Japan and the United States. The BOJ could purchase a broad range of assets, while it would be diﬃcult for the Fed to do so because the types of assets the Fed could purchase were tightly restricted.
The session chair, Esther L. George (Federal Reserve Bank of Kansas City), kicked oﬀ the general discussion by posing the question of whether asset purchases that could raise moral hazard issues should be conducted by a central bank or a ﬁscal authority. Ueda pointed out that the model provided a justiﬁcation for asset purchases, in particular the so-called QE1 conducted by the Fed. Regarding moral hazard issues raised by a central bank’s lender-of-last-resort function, Kuroda commented that it was not productive in the real world to try to completely eliminate the function in order to minimize moral hazard issues. Yoshino mentioned that a capital-injection scheme in which banks that had received injected capital must repay the capital could mitigate moral hazard issues. Regarding capital requirements, Adrian asked about a tradeoﬀ between the beneﬁts and costs. Obstfeld commented that one could use the presented model and conduct quantitative analysis on capital requirements from the standpoint of welfare. Kiyotaki replied that while imposing capital requirements could decrease 8 MONETARY AND ECONOMIC STUDIES / NOVEMBER 2014 Monetary Policy in a Post-Financial Crisis Era the volume of intermediation, capital requirements could produce a large welfare gain by reducing the fear of bank runs. Narayana R. Kocherlakota (Federal Reserve Bank of Minneapolis) argued that it might be inappropriate to rely heavily on the presented model for welfare analysis because the model seemed to lack an important mechanism in which a collapse in ﬁnancial activities spilled over into the real economy. Goodfriend argued that in the fall of 2008 the fear that had caused people to boost savings might cause the economy to collapse before the bank run kicked in. Végh pointed out that it would be diﬃcult to use the presented model to predict a crisis because a bank run in the model was unexpected. Hamada expressed his concern about whether it was appropriate to use a rational expectations model with an inﬁnite horizon to describe consumer behaviors within a ﬁnancial crisis.