The hidden impacts of rising interest rates
This notebook contains an explanation of the different methodologies we use in our Hidden impacts of rising interest rates report.
The key research in the report uses data from the World Bank's International Debt Statistics (IDS) database. With this data, we attempt to answer the following question: How much are countries expected to pay in interest on the money borrowed in a given year?
A note on the IDS data
The IDS data contains a wealth of information on external public and publicly guaranteed (PPG) debt. The different indicators can be viewed from a debtor perspective, and are broken down by the 'counterpart-area' (i.e. the creditor) to which different types of debt are owed.
The level of detail found in the data varies for each indicator. The IDS data is not presented on an individual loan basis. Instead, for interest payments for example, the data is broken down by year, debtor, creditor and by type of debt. That means we cannot tell how much interest is being paid on specific loans, but we have a single number for each type of debt/creditor (bilateral, multilateral, private banks, bonds).
Data on the terms of individual loans is unavailable. But data on maturities, grace periods, and interest rates are provided as weighted averages for a given creditor-debtor in a given year. The data is weighted by the amount lent by a specific creditor to a debtor in a given year.
On interest payments
In a given year, interest payments reflect the composition of the outstanding debt. Given different interest rates, maturities, grace periods, and borrowed amounts, the relationship between the interest rates and the interest payments in a given year isn't that straightforward.
However, we can use different data available through the IDS to estimate the expected interest payments for loans committed in a given year. For that, we use this data:
- Loan commitments (PPG): this data is in current dollars and reflects the total of all loans 'committed' by a given creditor to a given borrower in a given year.
- Average interest rate: this data is a weighted average of the interest rates of loans for a given debtor based on all the loans 'committed' by a given creditor in a given year. The data is weighted by the amounts committed.
- Average grace period: this data is a weighted average of the grace periods of loans for a given debtor based on all the loans 'committed' by a given creditor in a given year. The data is weighted by the amounts committed. For bonds, in most cases, the grace period is equal to the maturity of the bond (since the principal is only paid at maturity in most cases). We assume that interest payments are made during the grace period, and that principal payments are not.
- Average maturity: this data is a weighted average of the length maturities of loans (in years) for a given debtor based on all the loans 'committed' by a given creditor in a given year. The data is weighted by the amounts committed. We use this data in order to calculate the expected time during which the loans will generate interest payments.
A note on Net Present Value and our use of discount rates
Like the IMF in its Debt Sustainability Assessments, we use Net Present Value (NPV) in our calculations. This means we take each cash flow in the repayments and we discount it back to the 'present day' (the year of the loan commitment), using a discount rate.
NPV better reflects the "burden" of debt than the face value. For example, US$100 million payable in 10 years is less burdensome than a debt of US$100 million payable tomorrow. The NPV takes this time factor into account. When looking at loans with maturities spanning decades this becomes especially significant.
Calculating expected interest payments
To calculate the expected interest payments over the lifetime of loans or bonds we: - Use the amount committed and the interest rate provided. We assume that interest rates are fixed for the duration of the loan/bond. - Assume Equal Principal Payments with a single annual payment. - Assume yearly interest payments. We assume interest is paid during grace periods (but not principal), and that both interest and principal are paid after the grace period. - Assume there will not be defaults, rescheduling, restructuring or refinancing over the lifetime of the loans. - Use a discount rate of 5% to discount future cash flows. We do this to reflect the fact that money today (or at the time of commitment) is worth more than the same amount of money in the future.
We calculate the total interest due over the lifetime of loans (in NPV) in the following way:
We denote:
- C as the committed value or principal
- rate as the interest rate
- g as the grace period (if any)
- m as the maturity of the loan
- DiscountRate as the discount rate used
- y as the year of the payment
First, the yearly principal payments (in nominal terms) are calculated as
For the NPV of total interest payments, we sum the interest paid during and after the grace period. Note that for bonds, the grace period may be equal to the bond maturity.
The interest due during the grace period (in NPV terms) is calculated as follows:
The interest due after the grace period (in NPV terms), once principal payments are also being made, is calculated as:
And, finally, the total expected interest payments (in NPV of the year of the commitment) are calculated as:
Expected interest payments
The chart below shows how much a country/region is expected to pay in interest on the loans it agreed in a given year. These amounts are for the lifetime of the loan. You can also apply a discount rate—a rate of 0 means that the data stays in nominal terms.
Use the controls below to select a group of countries, creditor, a discount rate, and a 'new' interest rate for comparison.
Comparing the World Bank IBRD interest rates to Bonds
The chart below looks at the weighted average interest rates for Africa (and for 'other' non-African countries as a dropdown option). To produce these aggregates, the rates are weighted by the total loans 'committed' to a given debtor in a given year.
Health versus external debt service spending
This chart plots government health spending (x-axis) against interest payments on external debt (y-axis), both as a share of total spending.
The countries shown in purple, which are above the dotted line, spend more on debt service than they do on health.
The data is for 2020, based on health spending data from the WHO Global Health Expenditure database, and GDP and government expenditure data from the IMF World Economic Outlook database.
Many countries spend more servicing debt than on health
US Fed interest rate hikes by cycle
The chart below shows the pace at which the US has raised interest rates in the current cycle, and in other interest rate rising cycles in the past.
The lines track the amount by which the rate is changing, compared to the rate in place at the beginning of the cycle (in percentage points). The actual rate is shown on hover.
The x-axis tracks the number of weeks that have passed since the interest rate hiking cycle began.
Data sources
Interest cost analysis
The key research in the report uses data from the World Bank's International Debt Statistics (IDS) database.
We use the following indicators:
- Interest rates:
DT.INR.DPPG - Maturities:
DT.MAT.DPPG - Grace periods:
DT.GPA.DPPG - New loan commitments:
DT.COM.MLAT.CD(multilateral)DT.COM.PRT.CD(private)
FED rates analysis
The US FED rates chart uses data from The Federal Reserve Bank of Saint Louis, and is updated automatically as new data becomes available.
Health spending
The health versus debt spending chart uses data from the WHO Global Health Expenditure Database.
Additional data and replication code
For more information, as well as all code to replicate all of our analysis, please visit this report's GitHub repository.