Comment on “Fiscal Projections and Debt Sustainability in Japan”

Katagiri et al. (2025) project the time path of Japanese fiscal balances and evaluate the sustainability of Japan's public debt. Its unique strength is in its use of a macro-finance term structure model, which enables the authors to simulate the entire path of the yield curve, which incorporates endogenous time variations in term premia. This, combined with the projected maturity structure of the government debt, produces a forecast for interest payments that is deeply founded in the theory of finance. As correctly predicting the relative size between the interest rate and the economic growth rate is pivotal in this kind of exercise, I think the paper makes an invaluable contribution to the literature. I have four major comments and one minor one. First, I recommend that Katagiri et al. adopt the viewpoint of the consolidated government, namely, the general government combined with the central bank. Since the Quantitative and Qualitative Monetary Easing started in 2013, the Bank of Japan (BOJ) has purchased a substantial amount of Japanese Government Bonds (JGBs), especially 10-year bonds. Hence, they are not a part of the liabilities that the public sector as a whole owes to the private sector. At the same time, however, in exchange for those bonds, the BOJ has come to owe private banks almost an equal amount of liabilities in the form of bank reserves. Hence, the total amount of debt is little changed if we combine the government and the central bank. What changes markedly is its maturity structure. Bank reserves are extremely short-term liabilities. As monetary policy starts to tighten, the public sector as a whole would immediately have to start paying interest to the private sector. Furthermore, consider a risk scenario in which the BOJ has to raise its policy rate rapidly to combat inflationary pressure, and the yield curve inverts, as happened in the US. Damage to the fiscal balance, broadly defined, would be instantaneous. Treating the general government and the central bank separately would mask such a danger. Second, Katagiri et al.'s simulations assume that the general government will be able to more or less maintain its current debt maturity structure by rolling over the bonds that mature. But this may prove to be too optimistic. As the BOJ slows down its purchases of newly issued long JGBs, it is not clear if other entities will step up to replace it. There are already some signs that the government might have to increase its issuance of shorter-term JGBs. Third, Katagiri et al. uses the term structure model of Koeda and Wei (2023) which takes into account the presence of the effective lower bound on the short rate and incorporates the roles of macroeconomic factors. This is a very nice model but, in its essence, it is a no-arbitrage model. It is not clear if such a model is an adequate characterization of today's JGB market. I agree with Koeda and Ueno (2022) that, due to the presence of preferred habitat investors and a limited capacity of arbitragers, the BOJ has had an outsized influence on term premia. As it winds down its bond purchases and arbitragers return, term premia are expected to spike at some point. The question is when and by how much. I hope that Katagiri et al. in future will develop a new model of term structure along the line of Koeda and Ueno to answer this important question. Fourth, it would be informative to know the roles played by the macroeconomic factors in more detail. Katagiri et al.'s Figure A1 displays a projection for the future yield curve, but it is not clear how much of its variation over time is due to changing term premia. Would the results be similar in a simpler world in which the expectations hypothesis holds? If we find a sizable difference in the simulation outcomes, that would validate the usefulness of the authors' approach even more convincingly. My minor comment is on the non-monetary side of the study. Katagiri et al.'s baseline scenario assumes that the labor force participation rate increases “only” to 64.4% in 2040. However, there is much room for improvement. In 2023, the labor force participation rate for men over 65 years old was 34.8%, after a decade of improvement. In contrast, in 1968, it was 52.1% (presumably many of them were working as self-employed persons). Likewise, although the female labor force participation rate has increased, many women still work part-time. It is not impossible to expect that, with technological advancement and appropriate societal changes, the labor market participation of the elderly will be promoted further, and that more women will be encouraged to work full-time. Here, we can find hope (perhaps the only hope) that we might eventually be able to conquer the evil in the name of fiscal imbalances.