Latest Developments in the Canadian Economic Accounts
Guide to the Monthly Credit Aggregates

Acknowledgements

This work would not have been possible without the invaluable contributions and expertise of innumerable dedicated staff from the Bank of Canada and Statistics Canada.

Introduction

In the past, the Bank of Canada (the Bank) and Statistics Canada produced aggregate measures of borrowing, or credit, for sectors of the Canadian economy. The Statistics Canada measures were on a quarterly basis as part of the National Balance Sheet Accounts (NBSA) whereas the Bank of Canada published more frequent monthly statistics via their Monthly Credit Program (MCP). As part of a comprehensive credit data harmonization project, these measures were reconciled to better understand the statistical, conceptual, and methodological differences. At the same time, a joint project was underway between the Bank of Canada and Statistics Canada to develop a new collection vehicle for chartered banks initially called the joint balance sheet. This regulatory banking schedule would be administered by the Bank of Canada and designed in collaboration with Statistics Canada to serve the needs of both organizations and reconcile source data.

The outcomes of these two endeavors led to the decision for a full harmonization of credit statistics between the Bank and Statistics Canada with the agreement to move to a single set of credit statistics. Starting in December 2020, as part of the Monthly Credit Aggregates (MCA) program, Statistics Canada will be producing monthly estimates of selected liabilities that will be integrated into the overall macroeconomic accounts framework including the national balance sheet. This will provide users with a single, consistent, internationally comparable set of monthly estimates that align with officially recognized national accounting principles and classifications as detailed in the United Nations System of National Accounts Compilation Guide (SNA 2008).

This guide will serve as a brief review of the historical differences between each organization's credit aggregates, an overview of the conceptual and statistical changes that will occur as a result of the integration of monthly estimates within the NBSA, and a synopsis of the methods that will be employed to estimate the outstanding credit debt of households and non-financial private corporations by lending sector.

Reconciliation

The Bank and Statistics Canada conducted a joint reconciliation exercise to understand and identify key differences between their respective measures of household and business credit, including non-mortgage loans, residential and non-residential mortgages, debt securities, and equity. The outcome of this project included a series of articles covering the conceptual, methodological, and statistical differences for households and businesses and quarterly reconciliation tables that tracked these principal differences over time. While each organization incorporated similar data sources and portrayed the indebtedness of Canadian households and non-financial businesses in a similar light, various reconcilable differences existed due to the use of differing assumptions, classification systems, concepts, methodologies and coverage.

At the time of this exercise, the Statistics Canada measures were comprised of quarterly estimates that were published as part of the NBSA, which presents assets and liabilities between counterparties or in the case of credit, debtors and creditors, across all sectors of the economy including households and private non-financial corporations.  In other words, for each instrument, all holders of financial assets must have corresponding counterparties who hold the financial liabilities.

The Bank’s existing credit aggregates are presented based on the lender or creditor perspective, i.e., the holder of the financial assets where the counterparty is either a household or non-financial business. However, the Bank’s aggregates are more limited in coverage on both the lending sector and instrument dimensions (i.e., the Bank does not include all assets and liabilities for all sectors).

The table below summarizes the main differences that this exercise documented.

Concepts

The section below outlines the various concepts that will be implemented as part of the MCA program. Generally, the MCA measures the outstanding credit market debt of households and non-financial corporations based on lending from all sectors of the economy.

Valuation

Because the NBSA is on a debtor-creditor principle, it is important that the instrument liability of one sector align with the related asset of the counterparty sector. The outstanding stock of loans can be recorded on a gross or net basis, with the latter reflecting expected credit losses that may occur if a percentage of borrowers default and thus requiring a lender to write-off the value of those impaired loans. However, from the debtor’s perspective, no explicit assumptions are made for the likelihood of default. Thus, to ensure the debtor-creditor relationship is coherent and consistent, the gross amount of loans outstanding are recorded on both the liability side of borrowers and the asset side of lenders.

Also, like the NBSA, the MCA uses data for domestic lenders that is on a booked-in-Canada basis. This means that only lender activities from their Canadian operations are counted even though these activities may be with Canadians or non-residents. Similarly, borrowers are able to obtain funds from domestic or non-resident lenders. An implication of this treatment is that activities with non-residents are only recorded when they include a domestic counterparty and, consequently, transactions solely between non-residents are out-of-scope.

Outstanding stocks are recorded at the end of the reference period meaning the balance sheet will reflect a snapshot of all assets and liabilities at a single point in time. In this case, the NBSA captures a portrait of all assets and liabilities at the end of the calendar quarter while the MCA will record the portrait of credit debt on the last day of each month.

Lastly, outstanding debt is valued at market-value for loans liabilities and nominal-value for securities with the latter reflecting the debt from the viewpoint of the debtor that is owed to the creditor based on the terms of the contractual agreement. In the case of loans, the assumption is that market-value and nominal-value are equivalent and loans are not marked up on the asset side of the balance sheet.

Classifications and coverage

The NBSA adheres to internationally comparable taxonomies laid out by the SNA 2008 with some exceptions intended to reflect Canada-specific circumstances. To ensure coherence within the broader system of macroeconomic accounts, the MCA aligns to these same classifications.

Sector classifications

The Canadian System of Macroeconomic Accounts (CSMA) covers the entire economy and delineates economic activity across a number of domestic institutional sectors, the principal ones being households, non-profit institutions serving households, corporations, and government. The rest-of-the-world, while not a sector, but rather the same sectors identified above within other countries’ economies, is referred to as non-residents.

Debtor sectors: households and non-financial corporations

The MCA focuses on two of the major domestic sectors in its coverage of debtors (borrowers); that is, households and non-financial corporations. Households represent groups of individuals who share the same living accommodation, who pool some or all of their income and wealth and who consume certain types of goods and services collectively, mainly housing and food. An important inclusion within the household sector are unincorporated businesses that may engage in a variety of activities and incur debt. In the CSMA, unincorporated businesses are generally allocated to the household sector as these types of business cannot be reliably distinguished from their proprietors for the purpose of compiling sector statistics (i.e., frequently, the line is blurred between the assets of the household versus the business).

Corporations are classified to a variety of sub-sectors under both non-financial and financial corporations. Private non-financial corporations specifically are corporations not engaged in the production of financial services as their principal activity and which are not controlled by government (i.e., exclude government business enterprises). Additionally, the corporate sector will only include incorporated entities and some types of partnerships, but will exclude unincorporated businesses, which are classified to the household sector.

Creditor or lending sectors

The NBSA measures the stock of assets and liabilities across all sectors of the economy and presents these on a quarterly basis. While the MCA only focuses on two debtor sectors, from a creditor perspective, all lenders to these sectors are included. The table below presents the lender sectors whose activities are captured in the MCA as well as the corresponding groupings within the NBSA. The MCA is less detailed given the higher frequency estimates that are provided and the lack of a fully-balanced matrix with which to produce monthly estimates at the most granular sectoral level.

Financial vehicle corporations engaged in securitization

Each of the credit aggregates (mortgage and non-mortgage loans) include an adjustment for the effects of "securitization." Securitization occurs when loans are removed from the balance sheets of financial institutions and sold to special-purpose vehicles (SPV) who issue bonds, commercial paper, and other notes secured by the transferred assets. This credit, which is derecognized from the balance sheets of financial institutions, is recaptured in the credit aggregates through the inclusion of loans held by the SPVs referred to as “financial vehicle corporations engaged in securitization”. With the adoption of International Financial Reporting Standards (IFRS) beginning in January, 2011, financial institutions converting to IFRS now re-recognize securitized assets on their consolidated balance sheets when these are contained within an SPV that is non-arm’s length as per the provisions of the recognition/de-recognition rules laid out by the accounting standard. The most significant effect relates to the inclusion of securitized loans on financial institutions’ balance sheets, which were previously shown as loans held by SPVs including NHA mortgage-backed securities. These data are obtained from Statistics Canada’s Quarterly Survey of Securitized Receivables and Asset-backed Securities.

Instrument classifications

The focus of the MCA is on the credit market debt of households and non-financial corporations. For this reason, a specific subset of instruments are estimated on a monthly basis with the aggregate of these estimates providing a measure of total credit market debt for each of these sectors.

Non-mortgage loans

One important difference between the MCA and the NBSA involves the definition of non-mortgage loans to households. In the NBSA framework, the measure of household consumer credit includes only the financing related to current consumption, whereas financing for financial and capital investment expenditures is included in a separate category for non-mortgage loans other than consumer credit. Both categories combined cover the total non-mortgage debt of the household sector as defined in the MCA.

Mortgage loans

Generally, financial institutions will make a distinction between their mortgage assets based on the intended use of the mortgaged property, i.e., for residential or non-residential purposes. For example, if a residential care facility requires a mortgage, a financial lender will record this as a residential mortgage. The NBSA emphasizes the debtor-creditor relationship; that is, if the debtor is a corporation, then the liability will be classified as a residential mortgage in the non-financial corporations sector. Thus, residential mortgage liabilities may appear in various sectors of the economy. Similarly, non-residential mortgage liabilities are allocated to the various borrowing sectors with some non-residential mortgage borrowers classified to the household sector (i.e., unincorporated businesses).

Securities

The securities statistics cover issuances and holdings of financial negotiable instruments. Securities include debt instruments designed to be traded in financial markets, such as treasury bills, commercial paper and bonds, as well as equity instruments, such as listed shares. The statistics relate to outstanding amounts (stocks) and net transactions (new issues net of redemptions). Definitions and concepts are consistent with the recommendations of the Handbook on Securities Statistics, an internationally agreed upon framework for classifying securities instruments.

Data sources

The MCA incorporates a wide variety of data sources from administrative and regulatory sources, surveys, third-party data providers, and publically available information. These are available with differing frequencies and timeliness such that some portions of the estimates are more directly estimable on a monthly basis while others require additional assumptions. Moreover, the data sources for both the NBSA and MCA generally align given the high degree of overlap; however, the MCA may incorporate additional information to produce monthly trends that decompose the NBSA benchmarks.

Regulatory data on banks

Data on federally-regulated deposit taking institutions collected by the Bank of Canada, OSFI, and Statistics Canada are incorporated into both the MCA and NBSA programs. The balance sheets of chartered banks and the counterparties to whom they lend are compiled from this information.

The historical time series estimates are derived in part using information from the L4 schedule; this is one of the primary sources on which the Bank of Canada’s original monthly credit program is based.

A key banking schedule is the Z4—Balance Sheet by Booking Location, which provides monthly estimates of the balance sheets of chartered banks including selected counterparty information. This schedule was developed as a joint effort between the Bank of Canada and Statistics Canada as part of the credit harmonization initiative. It is administered by the Bank of Canada and collected through the Regulatory Reporting System.

In addition to the Z4, Statistics Canada also administers the AS – Non-Mortgage Loans Report, which elaborates on the balance sheet information provided by banks through the delineation of additional details on lending. The AS schedule is filed quarterly and, as such, is used as the basis for the quarterly NBSA estimates of loan assets for banks to which the MCA is reconciled. Finally, the MCP was based on information from the now discontinued L4 - Monthly average return of assets and liabilities.

Detailed cross-return rules exist between the Z4 and numerous other schedules to ensure consistent reporting and concepts. However, due to variation in definitions and objectives of the different returns some inconsistencies do exist. Firstly, the concept of loans to individuals for non-business purposes aligns across the Z4, AS, and L4. As such, the monthly Z4 and the quarterly AS should yield the same estimates and thus be used interchangeably; this identity is verified each applicable reporting period. Consequently, the monthly Z4 can be used to simultaneously construct both the monthly and quarterly estimates, which simplifies the methodology. However, there are differences between how the Z4 and AS capture loans to unincorporated business, which as mentioned earlier are classified to the household sector. While the AS has a clearly defined line item for reporting loans to unincorporated businesses, the Z4 is less precise. As a result, the AS is used as the quarterly benchmark, while the Z4 is a reliable proxy for deriving monthly trends. Lastly, banks provide information on loans to individuals for business purposes, a category that encompasses loans to corporations, non-profit institutions, government business enterprises, etc. The AS provides the most granular detail distinguishing loans to private non-financial corporations from all other counterparties whereas the Z4 is less clear-cut in this respect. As a result, the AS is used to establish the quarterly benchmark with the Z4 data used as a highly correlated monthly allocator.

It should be noted that, while some differences exist in definitions and specific instructions across the banking schedules, there is ongoing work to reconcile and align content where applicable. Additionally, banks produce their counterparty information using a variety of methods. Industry classifications recorded by the banks based on their clients’ business activities are one approach to classifying borrowers, but other approaches may be taken and the exercise is frequently dependent on the type of information collected at the branch-level, which varies between banks in level of detail and availability.

A comparison of the principal bank schedules used in the compilation of estimates is outlined in Table 4 below.

Non-bank data

Data for all other lending sectors except that of banks are outlined in Table 5. The data sources are varied as all other lending sectors are covered including non-bank financial and non-financial corporations, government, and non-residents. Moreover, there tends to be less detailed counter-party information among the non-bank lenders and full balance sheet information is generally available on only a quarterly basis. For this reason, various monthly indicators have been identified to be used in allocating the quarterly NSBA benchmarks to the monthly space.

Methodology

The overall methodology of the Monthly Credit Aggregates can be summarized as a decomposition of selected portions of the quarterly NBSA into monthly estimates. To accomplish this task, the MCA incorporates all available monthly data and ensures that this aligns as closely as possible with NBSA concepts. For example, with respect to chartered banks, the available banking information provides conceptually accurate information that is used to produce both quarterly and monthly estimates. The sections below detail the approaches taken across three domains: generating historical estimates, benchmarking, and forecasting when necessary.

Benchmarking and forecasting estimates

Given the important integration between the monthly program and balance sheet, the third month of each quarter is benchmarked to each quarterly release of the NBSA. Additionally, there will be a maximum of two months of estimates available as part of the MCA before the official release of the applicable NBSA quarter. Given this lag, these months will be forecasted using various methods (i.e., regression modelling) until the benchmarking process can take place, at which point all overlapping estimates will be harmonized and the subsequent months will again be forecasted.

Banks

For chartered banks lending, the monthly Z4 data provides information on loans to individuals for non-business purposes that is consistent with data received from the quarterly AS schedule. As such, the Z4 is used as a direct source when estimating household debt and is compared to the AS results each quarter to ensure cross-return coherence.  Information on loans to unincorporated business are only available on the AS, but the Z4 provides a reliable indicator for this amount and is consequently used to allocate and project the quarterly series.

The Z4 also provides information on lending to individuals and others for business purposes; however, this does not align as directly with what is collected quarterly from the AS.  The Z4 uses a broader definition of non-financial corporations than the AS, the latter being more consistent with the NSBA. As such, the Z4 monthly data is used as a close proxy to allocate the quarterly NBSA benchmark and forecast the months where the benchmark is unavailable. The differences between the Z4 and AS for non-financial corporations will continue to be investigated and the overall discrepancies are not expected to have a material impact on the monthly trends.

Non-banks

All non-bank lending sectors will be estimated monthly using all available indicators that serve as reliable proxies for monthly trends. These will be used in conjunction with ARIMA modelling to forecast months lacking a quarterly benchmark. If no reliable indicators exist, then linear interpolation will be applied to the quarterly benchmarks and forecasted values will be based on various methods, such as ARIMA models and year-over-year trends.

Timing of forecasting, benchmarking and revisions

The MCA will follow the same revision period as the NBSA and other accounts within the Canadian System of Macroeconomic Accounts. As a result, only certain months of the MCA will be open for revision at the time of each quarterly NBSA release. For example, at the time of the March release of the MCA, the first quarter NSBA estimates will not be available (see Table 7). Consequently, the three months of the first quarter will not be benchmarked to the NBSA.  At the time of the first quarter release of the NBSA, current quarter estimates will be available for the first time as well as revised estimates for the prior four quarters. The following MCA release for April will then incorporate the first quarter benchmarks as well as any revisions from the preceding four quarters; however, the April and subsequent reference months will not yet be benchmarked and will remain so until the second quarter release of the NBSA. However, there are two exceptions, the first being the two-months of the quarters that are not directly tied to the quarterly benchmarks and the seasonally adjusted estimates. Each of these may be revised without affecting the benchmarked values and such changes will be noted for data users.

Historical estimates

Selected monthly credit estimates from the Bank of Canada are available back as far as 1969 on a monthly basis whereas the NBSA quarterly estimates begin in the first quarter of 1990.

To produce a historical time series beginning in the first quarter of 1990, the Bank’s monthly estimates are benchmarked to the NBSA estimates using either a quadratic minimization (optimization) procedure when monthly data is available, or a linear interpolation process in the absence of monthly data.

Seasonal Adjustment

Generally, time series can be categorized into two different types: stock and flow. A stock series is a measure of the magnitude of attributes at a specific point in time, while a flow series refers to the accumulated value over a period of time. The MCA records the stock or value of assets and liabilities at the end of the quarter. The most practical approach to seasonally adjusting the monthly stocks is direct seasonal adjustment of the unadjusted MCA stocks. This approach was adopted for each component of the credit market debt. The resulting seasonally adjusted series were then reconciled to produce specific aggregate estimates including total credit market debt.

Appendix A
Concordance between the Bank of Canada’s and Statistics Canada’s monthly credit programs