Our ratings on financial institutions in Japan already factor in much of the impact any increase in interest rates in Japan and the U.S. would have. Still, we are closely watching those institutions that focus more on securities than loans for income. Their sensitivity to rates is much higher.

Here, we simulate the impact of an increase in interest rates. First, we look the effect a rise would have on net interest income (NII) on loans extended by Japanese banks. We assume Japanese banks we rate and those we do not are in similar situations. Second, we consider how a rise in interest rates would affect bond valuations in terms of unrealized losses, particularly for the 22 financial institutions we rate.

In addition, we make an international comparison of how the rate rises would play out. We look here at delta NII (changes in periodic NII from the previous fiscal year) and delta EVE (changes in EVE). The Basel Committee requires banks to disclose delta NII and delta EVE.

For Japanese banks, the changes in net loan interest income and unrealized losses on bondholdings have a more direct and substantial impact on banks' financial standings than changes in NII and EVE, which represent changes in the net incomes or value of all assets and liabilities.

Our assumption on the rise of the yen-denominated interest rates (+0.1%) differs from the one used for disclosures to the Basel Committee (+1.0%). Several factors are behind our assumption. These include the actual changes in yields on the Tokyo Overnight Average Rate (TONA), the Tokyo Interbank Offered Rate (TIBOR), and Japanese government bonds (JGBs), all of which have been small (see chart 1). In addition, the Bank of Japan has indicated it intends to keep its monetary policy accommodative for now.

Chart 1

Estimated Impact Of Rising Interest Rates

We forecast net loan interest income (loan interest income minus deposit interest expenses) would rise if rates increased. Such income would increase 2.6% in the first year and 6.7% in the third year, the final one under our simulation. We would expect most of the increase in net loan interest income to come from yen-denominated loans. We assume three years is the average life of the long-term fixed-rate loans held by financial institutions. For our simulation, we assume a third of the long-term fixed-rate loans on bank books are replaced by ones with higher interest rates (up +0.1% for those denominated in yen and +2.0% for those denominated in dollars) annually over three years. All estimates in this article are weighted averages, unless noted.

Regarding foreign currency lending, yields increased more than they did for loans denominated in yen. However, funding rates should rise by about the same amount as they do for foreign-currency denominated loans. This is because of the structure of the foreign currency funding used by Japanese banks and the short-term market situation for rates such as U.S. dollar LIBOR. Given the above, we believe the net interest margins (NIMs) for foreign currency lending are unlikely to improve unless improving loan quality through replacement occurs. Replacing low-margin loans to higher-margin ones without changing credit quality, for example, would improve loan quality.

Meanwhile, we also looked at unrealized losses on bondholdings if Japanese yen rates rise by 0.1% and U.S. dollar rates by 2.0%. Among the 22 Japanese financial institutions we rate, we think unrealized losses would amount to an increase equivalent to 12.8% of Tier 1 capital (as on March 31, 2022). We used the March figure as this is when financial institutions generally disclose detailed data on yen and foreign currency denominated bondholdings. In other quarters, institutions tend to disclose only limited data on such holdings.

There is a gulf between institutions with larger exposure to securities than loans and the rest. Those with the larger exposures (Norinchukin Bank, Shinkin Central Bank, and Japan Post Bank Co. Ltd.) would see a far greater impact than the major, regional, and other banks, all of which are more exposed to loans. Take, for example, the three megabanks--Mitsubishi UFJ Financial Group Inc., Mizuho Financial Group Inc., and Sumitomo Mitsui Financial Group Inc.--Japan's three global systematically important banks (G-SIBs). Our simulation shows the weighted average increase of unrealized losses on bondholdings to Tier1 capital write-downs is 10.3% for the three G-SIBs. The figure for the three institutions more exposed to securities is more than double, at 22.3%.

Table 1

Table 2

Comparison With European Banks

The Basel Committee requires banks to disclose delta NII and delta EVE, indicators of interest rate risk in the banking book (IRRBB). We compared the data of Japanese and European banks using the following formulas.

• Delta NII percentage: delta NII divided by fiscal 2021 net interest income;
• Delta EVE percentage: delta EVE divided by Tier 1 capital (core capital in the case of Japanese banks who apply domestic standard disclosure).

Assumptions in the Basel III disclosure data about increases in interest rates are 1.0% for yen rates and 2.0% for those in U.S. dollars. The assumption here for the increase in yen rates is higher than ours. In this comparison, positive delta NII equates to an increase in net interest income and a positive delta EVE means the economic value of a bank has decreased.

We found Japanese banks' delta NII and delta EVE percentages are slightly higher than the averages for European banks (see chart 2). These differences are attributable to balance sheet compositions of assets and liabilities and how the shape of the yield curves of respective currencies worked together.

From balance sheet structure, we see the main differences in the delta NII and delta EVE percentages between Japanese and European banks are derived from two main factors.

First, there are low loan-to-deposit ratios at Japanese banks. This, conversely, means there are high security investment-to-deposit ratios. In general, the average duration of bonds is longer, than that of loans, which increases interest rate sensitivity. Thus, high security investment-to-deposit ratios work to increase delta EVE.

Second, Japanese financial institutions have a high percentage of core deposits in their debt structures, which leads to limited needs for market-related financing. Consumers in Japan like to save, so deposit bases tend to be large and sticky. Therefore, when interest rates rise, the pace of interest tracking on the liability side is slower than on the asset side, which helps increase delta NII.

Taking a detailed look at the ratios of delta NII and delta EVE for ordinary Japanese commercial banks, we can see they are not substantially different from the weighted average of European banks. If we exclude the three financial institutions that have larger exposures to securities than loans (Norinchukin, Shinkin, Japan Post Bank), the divergence between European and Japanese institutions fades. In other words, this surfaces the extremely high interest rate sensitivity of the three investment-oriented financial institutions.

Chart 2

There are points to keep in mind when looking at Basel's IRRBB disclosure figures. First, since delta NII and delta EVE are net values that offset the situation on both sides of the asset and liability, they can differ significantly based on the assumptions each bank makes, namely in their calculations in the liability side. For example, variations in definitions of core deposits can lead to major differences in figures. There have been cases where revisions in definitions of core deposits have led to delta EVE changing significantly in Japan. Second, the 1.0% rise in yen interest rates used in disclosures might be much larger than any actual changes in market rates. The U.S. dollar rate (a 2.0% increase) is quite near to on-the-ground circumstances. However, the 1.0% rise in yen rates is large compared to the actual changes in TIBOR and yen-denominated bond rates. Given this, delta NII here indicates much larger increases in interest income than are actually happening. In addition, actual net interest income on loans of Japanese banks is increasing at a slower pace than delta NII here shows. This is because on the asset side, and particularly in terms of mortgage loans, Japanese loans are less sensitive to interest rates than European ones. European banks have already seen applied interest rates on mortgages increase; Japanese banks have not.

Implications

The impact of rising interest rates on the creditworthiness of rated individual banks will vary. It will depend on their underlying earnings capability and capitalization, as well as the risk profile of the interest rates on the bonds they hold. It is difficult to succinctly summarize the situation.

However, regarding the largest five banking groups, we believe that the downside risk of higher interest rates is largely factored into the current rating assessments, given their underlying earning capability and capital strength. The groups are the three G-SIBs and Resona Holdings Inc. and Sumitomo Mitsui Trust Holdings Inc. We see a possibility that growth in net loan interest income could outpace the results of our simulation. Higher yen lending rates and lower foreign currency funding rates could raise the pace here. Another positive factor could be that improvements in net loan interest income reaches the largest groups faster than they do regional banks. This is because the ratio of market-rate linked loans on their books is higher than at regional banks.

As for the 10 regional banks we rate, we expect net loan interest incomes to ultimately increase 9% by the 3rd year. This is slightly higher than the 7% over the same period at the majors. The difference comes from the composition of lending currency at regional banks. We assume that the improvement in net loan interest incomes will be greater in yen than in foreign currencies due to larger improvements in NIM. We expect higher funding costs in foreign currencies will mostly offset improvements in loan rates. However, overall domestic lending margins at regional banks might not improve as much as they do at major banks. This is because the competitive environment for retail lending (SME lending and housing loans) is severe. Also, the lower proportion of market-linked lending at regionals relative to major banks may hamper realization of the full benefits of rate rises. With regard to unrealized losses on bondholdings, we believe that risk factors have been fully incorporated into the current ratings on most regional banks we rate. However, there are a few regional banks where delta EVE to Tier1 capital is around 13%-14%. The 10 regional banks and groups we rate or analyze are Chiba Bank Ltd., Shizuoka Bank Ltd., Hachijuni Bank Ltd., Iyo Bank Ltd., Gunma bank Ltd., Bank of Kyoto Ltd., Hokkoku Bank Ltd., and Keiyo Bank Ltd., as well as Kyushu Financial Group Inc. and Hokuhoku Financial Group.

At the three securities-investment-oriented financial institutions, the two financial institutions are already assessed as having moderate risk positions. This indicates that each risk factor has been largely factored into the current ratings. However, the delta EVE and delta NII percentages are extremely high compared with those of ordinary commercial banks. This means they are highly sensitive to interest rates. Therefore, we are watching these institutions closely. In particular, we are focused on unrealized gains and losses on bonds and the possible side-effects of a decline in interest income because of the reduction of credit and interest rate risks associated with securities investments. In addition, we are closely monitoring future changes in market interest rates.

Related Research

• Asia-Pacific Financial Institutions Monitor 3Q 2022: Brace For Higher Credit Losses In 2023, Aug. 3, 2022
• Global Banking Outlook--Midyear 2022: Here Comes The Rain Again, July 20, 2022
• Banking Industry Country Risk Assessment: Japan, June 15, 2022
• Japanese Regional Banks: Widening Profitability Gaps Despite Favorable Fiscal 2021, June 20, 2022
• Japan Megabanks: Challenges Ahead For Sustained Earnings Capacity, June 14, 2022
• When Rates Rise: Not All European Banks are Equal, June 8, 2022