Forecasting Public Debt Charges for the Government of Canada – Focus on Nominal Bonds and Treasury Bills

Categories:

by Dominique Lapointe

Summary

  • Recent developments in global market interest rates have raised concerns over the fiscal impact of higher interest rates for the first time in almost ten years.
  • The Canadian yield curve has been rapidly reshaping in the last year. It is now both higher and flatter (meaning short-term borrowing rates have been rising faster than long-term).
  • The Institute of Fiscal Studies and Democracy (IFSD) expects a continuation of the tightening cycle by the Bank of Canada, thereby contributing to even higher interest rates in the next five years.
  • Higher rates, coupled with a debt management strategy focused on short- and medium-term bonds, should raise Treasury bills debt charges from $0.7 billion in the 2016-17 fiscal year to $4.0 billion in fiscal 2021-22. Nominal bonds charges should rise from $12.6 billion in fiscal 2016-17 to $18.3 billion in the 2021-22 fiscal year. Combined, those will represent 0.9% of GDP in 2021, from 0.5% today.

Latest developments

Recent movements in market interest rates worldwide are generating significant reactions and are contributing to financial market volatility. While investors and portfolio managers are wondering if these are the definitive signs of the end of the “global bond bull market,” the implications for policy makers and central banks are equally important. Indeed, if the recent trend continues, for the first time in almost ten years, governments will have to pay significantly more to finance their debt. Fiscal forecasts already assume higher interest rates. The magnitude of recent movements in global interest rates will, however, exacerbate this phenomenon.  

We interpret last year’s movements in Canadian domestic interest rates as two-fold. Because of very strong economic data starting in the second half of 2016, the Bank of Canada raised interest rates twice, in July and September. In effect, it sold the “insurance policy” it took against the oil price drop that started in 2014. Those rate hikes, coupled with market expectations of a faster tightening cycle, had a stronger effect at the front and middle end of the Canadian yield curve than at the back end.[1] This resulted in a flattening of the yield curve where the difference between yield-to-maturity on shorter-term bonds and longer-term bonds becomes smaller (label 1, Chart 1).

Chart 1: Canadian Yield Curve

Then came the New Year, a very strong employment report for the month of December and, consequently, another 25 basis points rate hike by the Bank of Canada on January 17th. The noticeable in the above chart since January 2018 is the almost parallel shift in the yield curve (label 2, Chart 1). Indeed, while the profile is still flatter than a year ago, at most tenures Canada’s yields-to-maturities have increased by approximately the same amount (20-25 basis points).

Going forward, these developments will have important and interrelated implications for Government of Canada public debt charges (PDC)––specifically, interest on debt (IOD). Below, we quantify those implications from a fiscal perspective by forecasting IOD on nominal bonds and Treasury bills, the most common types of debt used by the federal government to meet its borrowing needs. Indeed, in the last two fiscal years (2015-16 and 2016-17), IOD for nominal bonds and amortization on Treasury bill represented 55% of PDC, making up by far the largest PDC components.[2] In terms of volume, at the beginning of the last fiscal year, the outstanding amount of nominal bonds totalled $480 billion and the outstanding amount of Treasury bills was $137 billion. Together, nominal bonds and Treasury bills constitute 64% of total interest-bearing debt at the federal level.[3]  

Forecasting interest on debt

In order to forecast IOD, three sets of information have to combined:

  1. The interest rate at which future debt will be financed (a yield curve forecast);
  2. The funding requirements or the amount of debt that needs to be issued for refinancing needs, budgetary, and non-budgetary transactions; and
  3. The debt issuance strategy or the proportion of debt issued for each type of securities as well as across each tenure.  

Interest Rate Forecasts

First, while it is difficult–– or even impossible–– to forecast with precision the path of global market interest rates, some facts matter. For instance, the output gap in Canada is arguably closed, most measures of wage gains are increasing and, all else equal, so should inflation. The IFSD’s latest economic forecast called for the Bank of Canada to enter a gradual tightening cycle, with the target for the overnight rate reaching 2.75% by the end of 2021, 150 basis points higher than currently.  At the same time, longer rates should increase by a more moderate 90 basis points, to around 3.3% over that period due to a lower neutral real interest rate in the economy. This should result in a Canadian yield curve that is both flatter and higher than it is today (Chart 2).

Chart 2: IFSD Canadian Yield Curve Forecast

Funding Requirements

Second, to forecast the funding requirements, we proceed in multiple steps. To begin with, the amount of debt that needs to be refinanced every quarter is compiled, both for nominal bonds and Treasury bills, starting on April 1, 2017.[4] Those are the securities maturing in every quarter of reference. Then, the government’s fiscal policy is integrated using equation (1).[5]

In practice here, the IFSD’s forecast of the federal government primary account balance (total revenue minus program expenditures) is subtracted from gross refinancing needs.[6] This is because the Government of Canada is currently running a positive primary account balance. Finally, coupon payments and amortization on outstanding and maturing nominal bonds and Treasury bills are added to funding requirements.

Moreover, the government’s bond program is usually larger than the refinancing needs stemming from bonds maturing. One reason for this is the large cash management bond buyback (CMBB) program run by the Bank of Canada. To account for this, we scale up nominal bonds maturing in each quarter by an additional 30%. This is the average ratio of bonds bought back to bonds issued in 2017 (Chart 3).[7],[8]

Chart 3: Nominal Bonds Issuance Versus CMBB

Debt Management Strategy

Third, new debt is issued with respect to the debt management strategy on the last day of a specific quarter. For simplifying purposes, the share of debt issued in each type of securities (3-, 6-, and 12-month Treasury bills and 2-, 3-, 5-, 10-, and 30-year nominal bonds) is attributed through the “auction” of a single large nominal bond or Treasury bill each quarter.[9]

The debt management strategy used for the forecast period, 2018Q1 and onward, is consistent with the stated strategy in Budget 2017: “an increased focus on the issuance of short- and medium-term bonds (2-, 3 and 5-year maturities)”.[10] The historical debt management strategy, as well as the forecast used between 2018Q1 and 2022Q1, is presented in Chart 4.

Chart 4: Debt Management Strategy

While important changes in the debt management strategy can occur in the next five years, any radical departure from what has already been signaled by the Government of Canada cannot be foreseen. For instance, in our opinion, a relative cost-risk advantage exists by issuing more long-term bonds relative to short-term bonds due to long interest rates at a record low. This might be fading out in light of rising yields since the beginning of the year.

The average interest rate in each quarter forecasted as well as at each tenure is applied to each type of bond using the IFSD’s yield curve forecast. We assume nominal bonds will continue to be issued with a semi-annual periodicity. Treasury bills are issued as pure-discount products (no coupon bonds).

Interest on Debt

Interest on debt for both nominal bonds and Treasury bills is calculated at the end of each quarter. Public debt charges on nominal bonds are expected to increase from $12.5 billion in 2016-17 to $18.3 billion in 2021-22, a 47% increase. The predominance of 2-, 3-, and 5-year issuances to meet funding requirements weighs heavily throughout the 5-year forecast. This is because the spread between 2-year and 10-year bonds rates toward the end of our forecast is less than 20 basis points. Rollover costs on Treasury bills increase much faster: from $0.7 billion in 2016-17 to $4.0 billion at the end of the forecast horizon. Given that 66% of the debt management strategy consists of issuing Treasury bills with less than or equal to one-year tenure, this is hardly surprising. Indeed, rising interest rates affect Treasury bills and short-dated bonds more significantly than longer-dated bonds. Moreover, at the beginning of the last fiscal year forecasted–– that is, 2021Q2–– the costs of Treasury bills (as measured by the yield-to-maturity) are more than twice as expensive as they are right now (Chart 2).  

Chart 5: Interest On Debt

In terms of the stock of debt (outstanding amount), we forecast a steady rise in the amount of nominal bonds as a result of the debt management strategy putting greater emphasis on them. (That is to the detriment of Treasury bills stock.) However, starting in 2019-20, the value of Treasury bills in circulation should start climbing again, all else equal, due to the higher costs to refinance the whole stock of debt.

Chart 6: Outstanding Amount of Debt

In conclusion, future work will integrate real return bonds as part of the debt issuance strategy as they are indeed at the core of the Government’s debt management approach. Also, interest-bearing debt forecast will be completed with future pension and other employee benefits liabilities, thus rendering a complete picture of the repercussions of higher interest rates on the fiscal situation of the Government of Canada.

Disclaimer

FTSE TMX Global Debt Capital Markets Inc, FTSE International Limited, the London Stock Exchange Group companies or TSX INC. (“TSX” and together with FTDCM, FTSE and the Exchange, the “Licensor Parties”). The Licensor Parties make no warranty or representation whatsoever, expressly or impliedly, either as to the results to be obtained from the use of Canada yield curve (“the Index”) and/or the figures at which the said Index stands at any particular time on any particular day or otherwise. The Index is compiled and calculated by FTDCM and all copyright in the Index values and constituent lists vests in FTDCM. The Licensor Parties shall not be liable (whether in negligence or otherwise) to any person for any error in the Index and the Licensor Parties shall not be under any obligation to advise any person of any error therein. “TMX” is a trade mark of TSX Inc. and is used under licence. “FTSE®” is a trade mark of the FTSE International Limited in Canada and is used by FTDCM under licence.

 

[1] We define the front end of the yield curve as yields on securities with term-to-maturities up to but excluding 5 years. The belly of the curve ranges between 5 and (up to but excluding) 10 years, and the back end of the curve includes securities with maturities of 10 years and more.

[2] Other public debt charges (PDC) include, for example, premiums and discounts on real return bonds, foreign bonds, and retail debt and PDC on pension and other employees’ future benefits. For more details, see Table 3.8 of Volume 1 of the Public Accounts of Canada.

[3] On January 31, 2018, the outstanding amounts of nominal bonds and Treasury bills were $520 billion and $113 billion, respectively. See the Bank of Canada for more details.

[4] The source of data for the initial stock of nominal bonds and Treasury bills is FTSE TMX Global Debt Capital Market Inc. Those include stripped and reconstituted bonds but the latter both do not have a material impact on the forecast.  

[5] Bank of Canada. The Canadian Debt Strategy Model: An Overview of the Principal Elements (Staff Discussion Paper 2011-3). See Chapter 6. 

[6] Please see the latest IFSD fiscal forecast for more details.

[7] In reality, the CMBB program helps smooth out cash balances needed on key maturity dates. Generally, bonds bought back do not correspond to bonds being issued. The CMBB program targets bonds with a principal value of $8 billion and more and with a term-to-maturity of 18 months and fewer. See the 2016-17 Debt Management Report for more details.

[8] Additionally, the Bank of Canada undertakes bonds buyback on a cash and switch basis but those programs are not material to the forecast and therefore not integrated here.

[9] The Bank of Canada also runs a cash management bill program. While it can be substantial, the very short maturity of Treasury bills under this program (less than one month) does not materially affect their stock outstanding at any moment.

[10] See Annex 2 of Budget 2017 for more details.