1 of 13

JER-HIAS-Kakenhi Joint Conference

August 24, 2025

Iichiro Uesugi

IER, Hitotsubashi University

Comments on “Reversal of the BOJ’s balance sheet policy and liquidity dependence” by Professors Ugai and Osada

2 of 13

About the paper

Liquidity provision by the central bank and the impact on the banking sector during QE and QT

Examination of the relationship between

Reserves provided by the BOJ
Deposits (demandable and time) and credit lines provided by banks

Examination of prices for the liquidity provision as well

3 of 13

About the paper

Adopt the analytical framework employed for studying the monetary policy in the US (Acharya et al., 2024)

Examine the presence of liquidity dependence, the asymmetric bank behavior regarding the impact of QE and QT

QE: Commercial banks expand their BS by financing reserve holdings with claims on liquidity (demandable deposits and credit lines)
QT: Bank-issued claims may not shrink proportionately

4 of 13

How commercial banks change BS in response to QE

Purchases of bonds from commercial banks themselves do not expand their balance sheets.

Source: Acharya, Chauhan, Rajan, and Steffen (2024)

5 of 13

How commercial banks change BS in response to QE

Source: Acharya, Chauhan, Rajan, and Steffen (2024)

If the central bank purchases bonds from non-banks, BS of commercial banks expands.

6 of 13

How commercial banks change BS in response to QT

Source: Acharya, Chauhan, Rajan, and Steffen (2024)

During QT, if the central bank sells bonds to banks, BS of commercial banks will not contract. If the component of deposits remains the same, commercial banks are exposed to higher liquidity risk.

Panels B and C are an example of the “liquidity dependence”

7 of 13

Evidence for the liquidity dependence in Japan

In the aggregate estimations for deposits, there is some evidence for the liquidity dependence.
In the bank-level estimations for deposits, evidence is weaker.
In the aggregate estimation for credit lines, results are opposite to predictions

8 of 13

My overall comments

Interesting paper on a very important issue
The authors imply in the end that the liquidity dependence is not likely to cause liquidity crises in Japan due to the safety net.
However, before coming to this point, a further examination of the presence, magnitude, and mechanisms of the liquidity dependence may be important.
There are several things the authors may want to add.
In below are my suggestions.

9 of 13

Comment 1: Magnitude of the liquidity dependence

Though not so strong, there is evidence for the dependence.
Next question is how serious the dependence is for the financial stability.
One way to measure this is to compare the results with those for the US (Acharya et al, 2024)
US has experienced several liquidity crises in the financial market so it is meaningful to measure how close Japan is to the US.

March 2020 “Dash for cash” starting with corporations that drew heavily on lines of credit provided by commercial banks
March 2023 failure of Silicon Valley Bank among others

10 of 13

Size of estimated coefficients for Japan and the US

Although difficult to compare the liquidity dependence, results of aggregate estimations for the entire period are qualitatively similar and comparable
Size of coefficients: Japan << US

11 of 13

Why increase in bank deposits smaller in Japan?

Bond purchases by the BOJ may be mostly from banks, while in the US many of the purchases by the FED may be from non-banks.
Weak loan demand may be another factor

Small↑in loans leads to a small ↑ in deposits

This is consistent with the negative coefficients for credit lines

A larger rate of growth in reserves results in a smaller growth rate for credit lines.
Note that credit lines have been underdeveloped in Japan especially for SMEs due to legal constraints. (see Honda, 2022, JJIE)

12 of 13

Comment 2: Results using bank-level data

Another important issue is that the results using bank-level data are different from those using aggregate data.
More specifically, weaker evidence for the presence of liquidity dependence
However, due to heterogeneity across banks there may be some banks that are exposed to high liquidity risk

Banks that experienced a large increase in demandable deposits during the QE period and kept these deposits during the QT period
They may be vulnerable during a liquidity crisis

13 of 13

Comment 3: Consequence of the deposit insurance

The presence of deposit insurance in Japan, which is wider in coverage than the insurance in the US, may well reduce the probability of a liquidity crisis in the country.
However, it might have affected the deposit-taking behavior of banks.

Due to the safety system, banks may have an incentive to take more risk.
Small regional financial institutions that are not as heavily regulated as large banks may keep demandable deposits even during the QT period.