Watanabe (2021) starts with the question: “Why was the Kuroda bazooka unsuccessful?” “The BOJ has not achieved its target of 2%, and there is little prospect that it will be achieved in the near future.” Watanabe (2021) argues that there were two unforeseen events that prevented the Kuroda bazooka from working efficiently.

The first is coined as “Things that were expected to happen but did not happen.” They are factors that prevented inflation and inflation expectations rising. Household inflation expectations “rose slightly in the early days of the Kuroda bazooka regime, this did not last, and expectations have now returned more or less to the level before the Kuroda bazooka.” Also, Watanabe (2021) states that on the firm side, the pass-through was much lower than expected even with the Yen's depreciation.

The second is termed as “Things that no one anticipated to happen but that did happen.” That is “the large quantity of money supplied through the Kuroda bazooka itself may have had an adverse effect on the economy.”

In many conventional studies in monetary economics, including those on the zero lower bound of nominal interest rates, it is supposed regarding money-holding: where MU and OC denote the marginal utility and the opportunity cost, respectively. In such a case, to facilitate economic transactions most efficiently, the Friedman rule, namely, a zero nominal interest rate, becomes optimal. The fact that both the marginal utility and the opportunity cost are zero means that the utility from money holding is maximized.

As emphasized in Watanabe (2021), there are, however, non-negligible storage costs of holding money, which lead to negative interest rates as of now. With the storage costs, the new arbitrage condition is given by where SC denotes the marginal cost of storing money. Note that MU − SC can be negative and therefore, OC can be negative as well.

Watanabe (2021) claims that “[t]he presence of non-negligible storage costs of money has important welfare implications for the Kuroda bazooka.” With a non-negative storage cost, optimality is attained when MU = SC. Any increase in money beyond this optimal point is welfare-deteriorating.

Then, Watanabe (2021) estimates two money demand functions to compute MU and SC. With these inputs, Watanabe (2021) concludes that “the optimum quantity of money, measured by the ratio of M1 to nominal GDP, is 1.2. In contrast, the actual money-income ratio in the most recent quarter was 1.8.” Finally, Watanabe (2021) offers a policy prescription: “the Bank of Japan needs to reduce M1 by more than 30%”.

I enjoyed reading Watanabe (2021) and also learned a lot, in particular, how the existence of storage costs alters the optimal policy. It merits attention that Watanabe (2021) theoretically shows that a massive money injection can be harmful and gauges the welfare loss from it using parameters estimated from the data.

I have two comments, one related to the utility function and another related to simultaneity bias. Watanabe assumes a nonseparable utility function. There, the real money balance has direct impacts on consumption and inflation rates, although I believe that the degree of nonseparability should be small. Therefore, for the policy prescription to be readily useful, the benefit from money injection through this channel must be also evaluated. A money injection can increase aggregate demand. When aggregate demand is at a suboptimal level, this is the most important distortion to be remedied by monetary policy. It is not the suboptimality in utility from holding money.

The new insight offered by Watanabe that too much quantitative expansion can be harmful is very intriguing. This conclusion still remains valid even with a separable utility. Watanabe (2021) should be restructured so that its sole aim is to clarify the negative impacts from massive money injection. Academic contributions in this paper have little to do with the main question raised in the introduction: “Why was the Kuroda bazooka unsuccessful?” This question should be answered in terms of output and the inflation rate, not welfare. Higher welfare does not necessarily imply higher inflation or output gap.

Also, it would be informative if a robustness check is conducted not only for the storage cost functions but also for the utility functions.

Money demand functions are estimated independently. Watanabe writes that “how the demand for money responded to changes in the opportunity cost, money demand functions are estimated using data from 2006: Q2 onward.” This is true when the variables on the right hand side of Watanabe's (2021) equations (12) and (13), namely the opportunity costs, are exogenous. Equations (12) and (13) are, however, based on his equation (7), that is, one of the general equilibrium conditions. Therefore, money demand functions should be estimated in a system or by using appropriate instrumental variables, which affects real money balance not directly but indirectly only through the opportunity cost.

Watanabe ( 2021 ) 以这样一个问题开始：“为什么黑田火箭筒不成功？” “日本央行尚未实现其 2% 的目标，近期实现的可能性很小。” Watanabe ( 2021 ) 认为有两个不可预见的事件阻止了黑田火箭筒的有效工作。

第一个被创造为“预期会发生但没有发生的事情”。它们是阻止通胀和通胀预期上升的因素。家庭通胀预期“在黑田火箭筒政权的早期略有上升，这种情况没有持续下去，现在预期或多或少回到了黑田火箭筒之前的水平。” 此外，Watanabe ( 2021 ) 指出，在公司方面，即使日元贬值，传递率也远低于预期。

第二个被称为“没有人预料到会发生但确实发生的事情”。那就是“通过黑田火箭筒本身提供的大量资金可能对经济产生了不利影响。”

在货币经济学的许多传统研究中，包括关于名义利率的零下限的研究，假设关于货币持有：其中 MU 和 OC 分别表示边际效用和机会成本。在这种情况下，为了最有效地促进经济交易，弗里德曼规则，即零名义利率，变得最优。边际效用和机会成本都为零的事实意味着持有货币的效用最大化。

然而，正如 Watanabe ( 2021 )所强调的那样，持有货币的存储成本是不可忽略的，这导致了目前的负利率。有了存储成本，新的套利条件由其中 SC 表示存储货币的边际成本给出。请注意，MU - SC 可以为负，因此 OC 也可以为负。

Watanabe ( 2021 ) 声称“[t] 不可忽略的货币存储成本的存在对黑田火箭筒具有重要的福利影响。” 在非负存储成本的情况下，当 MU = SC 时达到最优。任何超过这个最佳点的货币增加都会导致福利恶化。

然后，Watanabe ( 2021 ) 估计了两个货币需求函数来计算 MU 和 SC。有了这些投入，渡边 ( 2021 ) 得出结论，“以 M1 与名义 GDP 的比率衡量的最佳货币数量是 1.2。相比之下，最近一个季度的实际货币收入比为 1.8。” 最后，渡边（2021）给出了一个政策处方：“日本央行需要将 M1 减少 30% 以上”。

我喜欢阅读 Watanabe ( 2021 )，也学到了很多东西，特别是存储成本的存在如何改变最优策略。值得注意的是，Watanabe ( 2021 ) 从理论上表明，大量注资可能是有害的，并使用从数据中估计的参数来衡量由此造成的福利损失。

我有两条评论，一条与效用函数有关，另一条与同时性偏差有关。Watanabe 假设一个不可分离的效用函数。在那里，实际货币余额对消费和通货膨胀率有直接影响，尽管我认为不可分离的程度应该很小。因此，要使政策处方有用，还必须评估通过该渠道注资的收益。注入资金可以增加总需求。当总需求处于次优水平时，这是货币政策需要纠正的最重要的扭曲。持有货币并不是效用的次优。

Watanabe 提供的新见解，即过多的数量扩张可能是有害的，这一点非常有趣。即使具有可分离的效用，该结论仍然有效。Watanabe ( 2021 ) 应该重组，以便其唯一目的是澄清大规模注资的负面影响。这篇论文的学术贡献与引言中提出的主要问题几乎没有关系：“为什么黑田火箭筒不成功？” 这个问题应该根据产出和通货膨胀率而不是福利来回答。更高的福利并不一定意味着更高的通胀或产出缺口。

此外，如果不仅对存储成本函数而且对效用函数进行稳健性检查，这将是有益的。

货币需求函数是独立估计的。Watanabe 写道：“货币需求如何响应机会成本的变化，货币需求函数是使用 2006 年第二季度以后的数据估算的。” 当 Watanabe ( 2021 ) 等式 (12) 和 (13)右侧的变量（即机会成本）是外生变量时，情况确实如此。然而，方程（12）和（13）是基于他的方程（7），即一般均衡条件之一。因此，货币需求函数应该在一个系统中或通过使用适当的工具变量来估计，它不是直接而是通过机会成本间接影响实际货币平衡。