Discussion of � �Cash Demand and �Demographic Changes in Japan��by Hiroshi Fujiki

JER-HIAS-Kakenhi Joint Conference�Macroeconomics and Japan’s Reality: Effects of Monetary and Fiscal Policies

​

August 24, 2025

​

Naoki Wakamori (Keio University)

Summary (1/3): Question/Data

• Q: How the aging and the diffusion of cashless payment technologies will impact future cash demand
• [+] Elderly people (larger share) use cash and hoard cash at home
• [-] Young people use less cash and hoard less cash at home

​

• Cohort based analysis is employed, �together with two data sources
1. Population projections (macro)
2. Online survey micro-data �(collected by the author)
• Cash on hand (COH)
• Cash at home (CAH)

Summary (2/3): Results

• Cash on hand will decline from 1.98T JPY (in 2020) to
• 0.95T JPY (≒ 1.5% annual decline)
• 0.73T JPY (≒ 2.4% annual decline) with technological progress
• Cash at home will also decline by 6T or 6.5T JPY (robustness are checked)
• As a 1% (2%) increase in the deposit rate would reduce CAH demand by 20% (25%), these increase further decrease cash demand

​

Summary (3/3): Implications

• Effects on the BOJ’s balance sheet and profit

​

• Annual declines of 2.4% in COH and 1.0% in CAH are translated an annual increase of 1.2T JPY in excess reserves
• When the BOJ pays 1% interest rate, the cost would rise by 0.012T, which is a 1% reduction in the BOJ’s profit

​

• A 1% (2%) increase in deposit interest rate leads to an increase in excess reserve of 23.44T (29.26T) JPY, resulting in a rise in annual costs of 0.35T (0.73T) JPY

​

• Changes in interest rate would be much larger than the changes in demographics

Overall: I like the paper

• What I like about the paper:
1. Combining macro data on population forecast and unique micro-data to obtain credible forecasts
• Micro data separates Cash on Hand and Cash at Home
• Careful treatment for censored data!!
2. Discussing not only cash demand projection, but how BoJ’s balance sheet and profits would be affected

​

• I have two major comments:
• Micro-foundation of technological progress and two-sidedness
• Exploiting micro data more?

​

​

​

Comment 1: Two-sidedness

• For modeling technological progress, the author assumes:

​

​

​

​

• My question: Is this assumption reasonable?
• Usage of cashless payment depends on merchants’ acceptance
• Due to two-sided nature of payments, as more consumers use cashless payments, more merchants accept them, which may generate non-linear effects
• Would the results change if such non-linear adoption dynamics were taken into account?

Comment 1: Two-sidedness

• My suggestions:

​

• Relate this number to the existing micro-payment literature, citing some papers that use empirical micro approach
• Huynh, Nicholls, and Shcherbakov (2019, BoCWP): Explaining the interplay between merchant acceptance and consumer adoption in two-sided markets for payment methods
• Li, McAndrews, and Wang (2020, JME): Two-sided market, R&D, and payments system evolution
• Yang and Ching (2013, MS): Dynamics of Consumer Adoption of Financial Innovation: The Case of ATM Cards

​

• Try some other numbers (and nonlinearity) for the technological progress parameter to check the robustness of the results, as the author did for the Parato coefficients

Comment 2: Exploiting Data More?

• The author says:

​

​

​

​

​

• So, I assume that quantifying the impact of diffusion of payment technologies is important
• Another way of addressing technological progress would be exploiting cross-sectional variation!

Comment 2: Exploiting Data More?

• At present, the projections are based only on age groups
• I initially thought that macro population projections were available only at the national level
• However, checking the IPSS website, I found that prefecture-level population projections are also available
• If the microdata contain respondents’ location, the author could incorporate prefecture-level projections to obtain more detailed forecasts
• Technological progress may have heterogeneous effects across prefectures and relying only on national-level data might underestimate these effects
• My suggestion: At least, checking micro-data to see any systematic differences across prefectures?

Comment 3: Importances/Implications

• The paper does not explain why cash demand projection is important. I would appreciate a brief discussion of its significance in the introduction.

​

• Possible areas of importance:
1. BOJ’s balance sheet management (already well covered!)
2. Monetary policy transmission
3. Payment infrastructure design
4. CBDC

If some of these are relevant, it would be useful to discuss them in Section 4.

Minor (Random) Comments

1. Exposition of the results: How about merging some tables and having some more informative figures?
• Is it possible to have prediction errors with this approach?
2. Adding some more information on micro-data!
3. Section 2’s title should be changed to Data and Methods
4. P6. The sentence starts with “Unfortunately, … “ should be rephrased
5. Population Projection for Japan (2023) is missing in ref.

​

​