In June 1990, the Child Support Project was announced by the federal, provincial and territorial ministers responsible for Justice. The federal/provincial/territorial Family Law Committee was given the mandate to study the issue of child support upon family breakdown in the context of actual costs of raising children in Canada.

In June 1991, the Child Support: Public Discussion Paper was released by the Family Law Committee. The Paper discussed current problems with child support, different alternatives for reform, and sought the public's views and comments on a range of issues relating to child support.

One of the options for reform discussed was that of child support guidelines. Although different types of guidelines were briefly described for information purposes, readers were not asked to offer their opinions on a preferred approach. Rather, they were encouraged to await the production of this research report which would focus on the financial aspects of child support and child support guidelines in particular.

This research report presents the results of the economic studies conducted to determine average expenditures on children in Canada. The data obtained from these economic studies was then applied to various child support guidelines and the results analyzed. In addition, the current income tax system and the tax implications of child support guidelines are examined. We recommend that this Research Report be read in close conjunction with the earlier Discussion Paper.

There is strong evidence that the costs of raising children are often underestimated by the parties themselves, the lawyers and the judiciary. An important first step was therefore to determine the actual costs of raising children in Canada.

The Family Law Committee sought proposals from a number of economists in different parts of the country on how the costs of raising children should be estimated. Three proposals were received and all three suggested that the optimum approach would be to estimate national expenditures on children using the most reliable database in Canada; the Statistics Canada Family Expenditures Data Base (FAMEX). All three economists preferred this approach to one of attempting to estimate the cost of the basic needs of children. Although data estimating the basic needs of children would have been very useful, national data does not exist. Given the costs of conducting such a study on a national basis, this approach was not seen as viable.

All three economists recognized that there is no perfect method of determining how much children cost. As a result, the approach chosen was to produce an estimate of average expenditures on children in intact families. One of the advantages of using an expenditure as opposed to a basic needs approach, is that it relates expenditures on children to the income levels of their parents. It also treats all children and families equally by purporting to allocate to all children the same percentage of parental income as the average children in an intact family would receive. However, some may have conceptual difficulties with this approach as a family which is no longer intact has neither combined income, nor combined expenses.

In relation to these issues, the reader is particularly referred to the June 1991 "Discussion Paper", more specifically on page 14 "Parental Income and Means - Needs of Children" and on page 39 "Extent and Limits of Child Support Guidelines: Advisory, Presumptive and Mandatory".

The Family Law Committee awarded a contract to each of three economists to estimate expenditures on children in Canada. Each economist recommended using a different economic model for estimating expenditures on children. The first chapter of the Research Report provides a detailed discussion of the four models selected by the economists.

Briefly, the four models are the Extended Engel, Blackorby/Donaldson, Adult Goods, and Consumption.

1. Extended Engel Model

The Extended Engel Model assumes that households that spend the same proportion of their total budget on necessities are equally well-off (i.e., food, clothing, shelter and household operations excluding child care). The "cost of a child" is interpreted as the income required to restore a couple with a child to their pre-child level of well-being. This is accomplished by finding the incomes at which a household with a child devotes the same share of total income to necessities, as an identical household without children. The difference in their income levels reflects the cost of the child.

2. Blackorby/Donaldson Model

The Blackorby/Donaldson Model is based on the assumption that there is an important link between behaviour and need: information about behaviour will allow us to develop information about needs. As well, in this model it is assumed that household "well-being" can be described using an index defined in relation to a set of necessary goods. Children and adults within the same household are also assumed to share the same level of material well-being. The "cost of a child" is the additional income required to make each member of a household with two adults plus one child, as well-off as each member of a two-adult household without children.

3. Adult Goods Model

The Adult Goods Model is similar to the Engel and Blackorby/Donaldson Model in that it looks for a good or set of goods that can be used as an indicator of well-being. It then compares the incomes of households sharing the same indicator to determine the difference between the income of the household with children and that without. Instead of household spending on food or on necessities, however, the Adult Goods model focuses on what parents spend on themselves as an indicator of well-being.

4. Consumption Model

Finally, the Consumption Model is grounded in what economic theory calls the life cycle hypothesis. Instead of finding ways of comparing the "well-being" of families with and without children (like the Engel, Blackorby/Donaldson and Adult Goods Model), the objective of the Consumption Model is to follow the same household through its life cycle to see their expenditures change with changes in demographic structure. The difference between expenditures in periods with children and without children is taken to be the cost of the children.

Although four models were used to determine expenditures on children, the following common patterns emerged from the results:

1. A family's income level determines the level of its spending on children.

2. Expenditures on children increase with the level of family income. However, the increase is not necessarily in proportion to the income. In the Blackorby/ Donaldson model, there is a proportional difference between lower and higher income families that increases with the age of a child. It is suggested that this may be because such households have more possibilities for expenditure on older children. The Adult Goods model yields the opposite result. It suggests that upper income families spend more dollars but a smaller proportion of their income on children than do lower income families. In the Consumption model, there is no particular pattern to the proportion of income spent on children. For a family with one young child this proportion rises with income; for a family with an only child aged 12-18 it is more or less constant and for a family with two older children it falls with income.

3. All models suggest that expenditures on children increase with the number of children. However, expenditures do not increase proportionately. In other words, there are economies of scale realized in having larger families. In the Adult Goods model for example, upper income families spend 30-35 per cent less on a second child than on the first, about 50 per cent less on the third than on the second, and an additional 50 per cent less on the fourth than on the third.

4. The economic models revealed only slight regional variations in the overall levels of expenditures on children. According to the Adult Goods model, families living in Alberta spend the most on their children, both in terms of dollars spent and as a percentage of income. According to the Consumption model, the region of residence has a slight impact on expenditures on children. Finally, according to Blackorby/Donaldson and Extended Engel models, the level of urbanization and region of residence does not affect the proportion of income spent on children. However, in actual dollars, families living in Ontario spend more on their children than families living in the Atlantic provinces due to the difference in the average income of parents in these two regions.

The following chart compares the average percentage of gross income spent on children by number of children in the family, in all income categories as estimated using each of the models. It must be noted that day care costs and income tax implications are not included in the estimates provided by the economists.

The second chapter provides a brief description of seven child support guidelines and demonstrates, using six specific fact scenarios, how these guidelines operate when combined with the different economic models. Approximate child care costs and income tax implications are also considered within these scenarios.

The selected child support guidelines are:

Income-Shares

Income-Shares with Reserve

Delaware-Melson

Flat Percentage

Flat Percentage with Reserve

Australian

Income Equalization

It is important to note that the amounts chosen for the various components of the guidelines were selected for illustration purposes only. For example, for the purposes of some of the guidelines, an amount had to be chosen for the parental self-support reserve which is a basic amount parents are allowed before they are required to pay child support. Should guidelines, in fact, be implemented, different amounts could be selected based on the results of the consultation process.

Chapter three examines the strengths and weaknesses of the different economic models for determining expenditures on children and indicates how these amounts would be applied within child support guidelines. The analysis was based on simulations performed on the four economic models, in combination with the various child support guidelines, using selected family income levels. The tax implications and child care expenses were not considered in the economic models in order to allow us to determine the effect of the tax and child care expenses on a case by case basis. Child care costs were also added after the application of the guideline for the following reasons: these costs are much higher in the early years, they can easily be established and they vary considerably from family to family. The results of the economic models and different guidelines were evaluated with reference to the general objectives of child support as developed by the Family Law Committee.

The evaluation criteria were as follows:

1) Consistency: whether there are dramatic increases in child support payments with only slight increases in the total gross incomes;

2) Robustness: whether the model or guideline can accommodate a wide spectrum of combinations of incomes and children;

3) Absence of extreme values: whether the results contain extremes or not, and if extreme results are obtained, are they unreliable.

The results of the evaluation of the economic models can be summarized as follows:

(i) The Extended Engel Model:

The estimated expenditures generated by this model are the lowest of the four economic models. This is particularly true for families with large number of children. Overall, this model is simple, produces numbers which are consistent, robust and does not require additional amounts of data. However, this model often yields the lowest values.

(ii) The Adult Goods Model:

The expenditures on children using this model are the second highest of the four models. This model produces the highest child support awards for families with low family income but in high income families, this model produces the lowest child support awards where the number of children is small and particularly where the children are older. This model is simple, produces numbers which are consistent, robust and does not require additional amounts of data.

(iii) The Consumption Model:

Of the four economic models, the consumption model generally produces the second lowest expenditures on children and produces the lowest child support awards in low income families, particularly where the number of children is small. This model is simple, produces numbers which are consistent and does not require additional amounts of data. However, this model is not seen as being robust.

(iv) The Blackorby/Donaldson Model:

The estimated expenditures on children produced by this model are generally the highest of the four economic models. This is particularly true for families with high income. In cases where there are four children, the child support awards based on this model can be as high as 50 per cent of the gross income of the non-custodial parent. The high estimates arise mainly as a result of the apportioning of the housing costs since all members of the family are assumed to equally share the total housing costs. Overall, this model produces numbers which are consistent and robust. This model requires more data than is available in the FAMEX database and is theoretically complex.

The analysis of the following child support guidelines did not take into consideration the tax implications or child care costs. The results can be summarized as follows:

(i) Income Shares

This guideline yields child support awards which are proportionate to the income of the two parents. It is straightforward and easy to explain. With the exception of the Flat Percentage Guideline, this calculation is the easiest to apply. Overall, this guideline is simple, consistent, robust, does not yield extreme values and does not require additional amounts of data.

(ii) The Income-Shares with Reserve:

This guideline yields child support awards similar to the Income Shares Guideline. The only difference is that the self support reserve allows both parents to cover their basic needs before paying child support. Therefore, the parent with the higher income will be responsible for a larger share of the expenditures on children than he or she would have under the simple Income Shares Model. As the amount of the self-support reserve increases, the parent with the higher income will pay a smaller percentage of the total child support award.

One concern relating to this guideline is that the percentages arrived at using the economic models are based on gross income and not income after the deduction of a parental self-support reserve. The amount of the child support award would not, therefore, correspond to the amount produced by the economic models. However, this guideline is consistent, robust and does not yield extreme values. This guideline is slightly more complicated than the Income Shares guideline and requires a determination of the value of the self support reserve.

(iii) The Delaware (Melson):

This guideline yields the highest child support awards. It combines two components in the awards: sharing the basic needs of the children and a specific percentage of parental income as a cost of living adjustment. If a high dollar amount is attributed to the child's basic needs, the child support awards can be higher than the total estimated expenditures on the children. This is because a percentage of the parental income is added to the child's basic needs.

Overall, this guideline is consistent but not robust, yielding extreme values, and does require additional information on the value of the self support reserve, the child's basic needs and the average percentage of gross family income spent on children. In addition, while this guideline may be theoretically sound, the additional steps make it slightly more complex.

(iv) The Flat Percentage guideline:

This guideline will generally yield child support awards which are similar to those produced using the Income Shares guideline. Overall, this guideline is the most straightforward of all and is the easiest guideline to apply in terms of the amount of calculation required. It is also consistent, robust, does not yield extreme values and does not require additional amounts of data.

(v) Flat Percentage with Reserve guideline:

This guideline will in most cases yield the second lowest child support awards. Overall, this guideline is consistent and robust. It is simple to apply although slightly more complicated than the Flat Percentage guideline. However, this guideline yields relatively low child support awards and requires additional information on the specific amount of the self support reserve. In addition, one concern relating to this guideline is that the percentages from the economic models are derived from gross income and consequently should be applied to gross income only and not to income after a deduction of a self-support reserve. As a result, the total amount for the child's needs as calculated using this guideline would be different from the results of the economic models.

(vi) Australian guidelines:

For the purpose of our report, formula #3 was selected. This formula applies in cases where custodial parents' gross income is above average weekly earnings which are determined at $22,112 in 1991. This guideline will generally yield the lowest child support awards. If the income of the custodial parent is lower than $22,112, a different formula will apply and will yield results which are the same as the flat percentage with reserve guideline. If the custodial parent earns more than the non-custodial parent, the child support awards will be even lower. On the other hand, if the non-custodial parent earns more than the custodial parent, the child support awards will be similar to those produced under the Income Shares guidelines. This guideline is the most complicated of all, since it requires a choice to be made between different guidelines depending on the level of income of the custodial parent. In addition, this guideline is not robust (as it may occasionally yield negative child support awards). This guideline also requires more additional information than any other guideline.

(vii) The Income Equalization:

This guideline will yield the highest estimated support awards if the non-custodial parent has a higher income than the custodial parent. However, this is only true if the non-custodial parent does not have subsequent dependents. If the non-custodial parent has dependents, then the estimated awards may be much lower than under other guidelines. In some instances the awards may even assume negative values.

Overall the guideline is not simple and may be more difficult to explain. It is not consistent, as the calculated support awards vary widely with different income profiles. It is not robust, yielding extreme values, and it requires additional dollar information for the cost of living components.

Income tax implications have a significant impact on child support awards. Chapter four presents a brief overview of the Canadian income tax system as it pertains to child support, the policy behind the system and the effects of such a system on child support payments.

The chapter describes how child support payments are deductible to the payor and must be included as income of the recipient. According to the Department of Finance, the policy behind such a system is based upon four major principles:

1. Where a deduction is received for a payment by the payor, the recipient should pay income tax on that payment.

2. Tax payers with similar incomes from different sources should pay the same amount of tax.

3. The tax assistance offered by the deduction may provide an incentive for the payer to make the payments.

4. The benefit of the tax subsidy to the payer should be passed onto the child in the form of higher support payments.

Different tax credits and deductions may apply to custodial parents depending on several factors. The following credits and deductions and the conditions for their application are discussed: the equivalent-to-married credit, the deduction of child care expenses, the refundable child tax credit, the refundable goods and services tax credit, the deduction of medical expenses for children and tuition and education credits.

The effects of the income tax system on child support awards are then discussed. In determining child support awards under the current system, taking the income tax implications into account is of utmost importance for the custodial parent and the child. Neither, the Divorce Act nor provincial and territorial legislation governing support specify how and when these tax implications should be calculated, or if and how the deduction should be shared between the parents in order to benefit the child.

Another issue which is raised is the requirement that custodial parents include the amount of support in their income and the strain this may cause for them when the amount was insufficient to start with. In some circumstances, the inclusion of support payments result in a change in a parent's marginal income tax rate. This may make a difference in the value of the support order and variation of the order may be justified. As well, where child support awards are determined by way of negotiations and settlements, there may be no way to ensure that the custodial parent is compensated for the income tax that he or she will have to pay on the amount received.

Finally, if child support guidelines are adopted, the income tax implications must be dealt with and proper income tax calculations must be effected. Direction should be given as to how to complete such calculations and how the benefit of the deduction should be shared in order to assist the child. However, such calculations may complicate the application of child support guidelines which sought to simplify the process of support determination as well as provide more fairness in child support awards.

The Family Law Committee hopes to receive comments on the two public discussion papers.

Readers may wish to comment only on the issues of substance which were raised in the Child Support: Public Discussion Paper released in June, 1991. On the other hand, readers may also choose to comment on the different economic models presented, on the different types of child support guidelines, on the effects of combining the different economic models with the different guidelines and on the income tax implications.

Written submissions may be sent to:

Child Support Project

Federal/Provincial/Territorial Family Law Committee

c/o Department of Justice

Room 758

239 Wellington Street

Ottawa, Ontario

K1A 0H8

and to your own provincial or territorial Attorney General's Department or Department of Justice.

Submissions should be received no later than October 11, 1992.