The
Economic Freedom
Network

Appendix A: methodology and data

This report was written for the non-specialist reader who does not have an extensive accounting or financial background. For those who require more detail, a technical discussion of the methodology is included here. The following section summarizes the methods used and sources referenced to calculate the Total Government Liability figures and the international comparison.

Methodology for computing total government liability for Canada

Unfortunately, government reporting of fiscal performance lags behind the events. As a result, totals for the majority of the liability categories had to be estimated for 1996 and 1997. The basic projection technique was to extend the trend of the most recently available information. In order to apply an unbiased and fair rate of growth, an average rate of change was calculated using the four most recent years of actual data. The average rate of increase or decrease was then applied to the most recent year to estimate 1997 values. For instance, figures for provincial public debt guarantees are only available up to 1995 from Statistics Canada's Financial Management System (FMS). To calculate the 1996 values, the 1995 figures were multiplied by the 1991 to 1995 growth rate. These 1996 values were then also multiplied by the 1991 to 1995 growth rate to arrive at the 1997 figures. In most cases, this general formula was applied to calculate 1997 values. In situations where four years of data were not available, data from the most recent fiscal year was used.

A different process was developed to determine the direct debt of the federal and provincial governments. The ratio between the figures from the FMS for direct debt and the figures from the Public Accounts (PA) for direct debt was found using three years of actual data. To generate an estimate of the FMS figures for debt for 1996, this FMS to PA ratio was applied to the 1996 figures for debt from the Public Accounts. The 1997 estimate was calculated by applying this relationship to the direct debt figure indicated within each 1997 budget; the 1998 budget was used if it was available. The resulting estimates reflect policy changes that occurred during the interim reporting period where no FMS information is available. Incorporating recent budget estimates of direct debt is essential in arriving at a reasonable estimate. For instance, the methodology based on the 1991 to 1995 growth rate estimates Alberta's direct debt at over $3.7 billion; the revised methodology estimates the direct debt at $1.7 billion.

Data for Canada

The majority of the Canadian data in this report is from Statistics Canada's Financial Management System and the provincial and federal public accounts. Table A1 is a list of the sources for the Canadian data by category, giving the most recent reporting date for the various categories of liabilities

In addition, The Fraser Institute commissioned a study from Statistics Canada and another from the Office of the Superintendent of Financial Institutions (OSFI).

The Public Institutions Division of Statistics Canada provided a study detailing financial performance and debt levels of tax-supported and self-supporting federal government business enterprises.

The Actuarial Services Division of the OSFI provided a comprehensive study that estimated the unfunded liabilities of various government programs including the Canada Pension Plan, Old Age Security (OAS), and the health-care system. The estimates are based on conservative assumptions: inflation at 3.5 percent, earnings growth at 4.5 percent, and the interest rate at 6 percent.

There are a variety of methods that could be used to allocate federal liabilities, such as income per person, population, or some taxation-based measure. This study uses the provincial contribution to federal tax revenues because this best reflects the distribution of the federal debt burden. Applying federal liabilities this way generates different liability values for each province, a procedure that acknowledges and captures broad regional deviations. The calculations of tax shares encompass all federally mandated taxes, both direct and indirect. A 5-year average of the federal tax-share statistics is applied to each category of federal liabilities to derive each province's share. To maintain consistency, this 5-year average is applied to the historical federal liability figures.

The methodology is modified for the Canada and Quebec Pension Plans. Quebec is allotted the full value of the Quebec Pension Plan's assets, liabilities, and unfunded liabilities. The contributions of each province and of the two territories to the Canada Pension Plan are used to distribute the CPP's assets, liabilities, and unfunded liabilities.

International Comparison

A total of 151 countries from all income groups were analyzed to determine an estimate of subsistence income, the minimum amount of spending required to meet a level of consumption that sustains life. Comparing consumption and income patterns across the entire spectrum of nations found that even if a country generated no income, it would still consume $183 US per person per year.

The methodology for developing the severely indebted country (SIC) list consists of the following three steps.

For the countries where information is available, determine the debt per person.

Calculate the income per person above subsistence level in each country.

Compute the ratio of debt per person to income per person.

Five separate income groups were created for this analysis: (1) high income (OECD nations), (2) upper-middle income ($2,305 to $9,312 US per person a year), (3) lower-middle income ($726 to $2,304 US per person a year), (4) upper-low income ($349 to $725 US per person a year) and (5) low income (under $349 US per person a year). The World Bank income categories were used to generate the three main classifications and the medians of the low and middle income groups were used to create the additional divisions. The upper limit of the Upper Middle Income group was increased to include Slovenia. Appendix D contains the country lists and their rankings.

Click Here to Appendix D

The World Bank statistics, which are used for all of the middle and lower income countries, assess only external debt and, therefore, underestimate the debt of countries like Brazil and Argentina, who have the capacity to generate internal debt through domestic savings. Nevertheless, extensive research performed for previous editions of this study found that, in most cases, external debt is a useful approximation of total debt.

The debt of the high-income nations is also underestimated. The system of national accounts employed by the World Bank and the OECD to estimate the debt of OECD nations excludes a calculation of the unfunded liability of public pensions. Generally speaking, OECD nations provide generous public pension plans. Calculating debt including unfunded pension liabilities, as we did in the detailed section on Canada, would substantially increase the ratio of debt to income and lower the rankings of most of the OECD nations. Since few middle and low income countries have the financial ability to provide public pensions, their debt is not distorted to the same extent.