For most families, the birth of a child is a happy and positive experience, but the addition of a newborn baby to a household may also, in some ways, decrease parental physical and psychological wellbeing. There are many well-known factors that can increase pressure on a family when a newborn arrives, such as changes in the hours of unpaid and paid work and role adjustments. Of particular interest in this paper is how an "extra mouth to feed" can increase pressure on a family's financial resources.¹ While any increased burden on the household budget is likely to be an issue for affected families, it is also of key interest to policy-makers who are looking at ways of assisting these families to make a smooth transition through pregnancy and childbirth. While this paper does not attempt to evaluate any particular policy, research that contributes broadly to our understanding of family financial pressures and how expectant and new parents manage these financial demands is particularly important in the current policy climate, due to the recent significant changes to payments centred around the birth of a child.² In this study, we also focus on expenditure on electronic goods. This is because various media reports suggest that some families spend these payments on adult consumer products rather than on goods and services for the benefit of the newborn (see, for example, Haynes, 2008).

International and Australian studies have examined the costs of children more generally (for example, see Henman, Percival, Harding, & Gray, 2007), but until recently there has been little research that specifically focuses on how newborns affect the household budget. In 2010, the authors used data from waves 6 and 7 of the Household, Income and Labour Dynamics of Australia (HILDA) survey to examine how the arrival of first-born, second-born, and third- or subsequent-born babies affect household budgets for stable couple families (Brandrup & Mance, 2010). The current study extends the authors' prior work by including wave 8 of the HILDA survey. The inclusion of these data has allowed the authors to refine the analysis with an increased sample size, but also created the opportunity to conduct a "skipped year" analysis, whereby spending patterns of couples with newborns can be compared with their spending patterns before the study women were pregnant.

Does a newborn increase family expenditure?

The constraining effect of income

Intuitively, it might be expected that there would be a straightforward relationship between the number of consumers in a household and their expenditure, but prior studies indicate a range of family attributes that affect the expenditure patterns of families with children. While adding consumers to a household generally increases family expenditure, individual tastes, preferences and the amount of money that parents have available to spend on their children are also key influencing factors (McDonald, 1990). As income generally constrains expenditure, for a constant level of income, the costs of goods and services for extra children will need to be offset by reductions in expenditure on other household items (McDonald, 1990; Percival & Harding, 2005); although households may be able to use credit to some extent to smooth out expenditure patterns. As such, we would expect that when adults become parents, they would reallocate their financial resources to provide adequate food, clothing, shelter and education to their children. If demands on household expenditure are great, necessities would consume a higher proportion of the available resources (for a fuller discussion, see Brandrup & Mance, 2010).

When a newborn child enters a household, the relationship between pre- and post-birth expenditure and income may not be as straightforward as reallocating resources to meet essential baby-related expenditure demands. For instance, there may be changes in total household income coinciding with the birth of a child that affect the amount of money available to spend. Income may decrease temporarily or permanently as mothers withdraw from the workforce or, on the other hand, increase because of rises in father's earnings due to career advancement or longer working hours (for example, see Craig, 2002). Alternatively, household income may increase due to becoming eligible for family assistance.

The influence of demographic factors

Along with household income, prior studies have found demographic factors that influence how families allocate their financial resources (Exter, 1992; Percival & Harding, 2005; Valenzuela, 1999). These factors are covered in more detail in the authors' prior work (Brandrup & Mance, 2010), but for completeness, an overview is briefly provided here.

Expenditure on food has been associated with family size, employment status, marital status, ethnicity and consumption of takeaway food (Fan & Zuiker, 1998; Kaushal, Gao, & Waldfogel, 2007; Lee & Huh, 2004; Passero, 1996; Valenzuela, 1999). On average, adding children to a household incrementally increases expenditure on food, as economies of scale do not apply to food in the same way as they might apply to other expenditure items (Valenzuela, 1999). At the same time, there appears to be an upper limit on the amount of groceries a family is able to consume (Percival & Harding, 2005).

In contrast, decreasing average expenditure on children's clothing has been observed with increasing numbers of children in the household (Valenzuela, 1999). For the first child, significant adjustments in expenditure on commodity items are required. For subsequent children, recycling of resources operate as clothing is passed down from older to younger children. Spending on clothes has also been found to increase with labour force participation and income (Kaushal et al., 2007; Percival & Harding, 2005) and, with increasing wealth, a higher proportion of clothing expenditure is found to be purchases of "comfort" goods rather than necessities. Carlucci and Zelli (1998) argued that although a basic level of clothing may be thought of as essential, wealthy families have more scope to increase their expenditure on clothes as a luxury item than on basic necessities such as groceries because, as stated previously, there is an upper limit on how much food an individual can consume.

Not surprisingly, the addition of a young child to a family may increase demand for child care. In prior studies, increasing costs of child care have been found to be associated with increasing family income, mother's employment hours, the number of children in care, service accessibility, the availability of relatives and partners, and service cost (Doiron & Kalb, 2005; Walker & Reschke, 2004). In contrast, there is unlikely to be additional demand for education expenditure associated with young children until they attain school age (Exter, 1992).

In terms of health, cross-sectional studies have found expenditure to decrease with the presence of children (Exter, 1992; Valenzuela, 1999). However, the results of Exter's study were based on comparisons between all couples with and without children, thus the higher health expenditure for childless couples may reflect their older average ages in this study. In contrast, in Brandrup and Mance (2010), increased health expenditure was associated with the arrival of newborn children, regardless of whether they were first-born or subsequent births.

The cost of transport may also impose a significant burden on the household budget as a family increases in size (Valenzuela, 1999), but this depends on the extent to which the household uses private or public transport. Car-owning households have been found to spend substantially more on transport due to the purchase price of the car and ongoing expenses (Saunders et al., 1998); however, average spending on transport also tends to increase as children age, and with parents' increasing labour force attachment and life-cycle stage (Australian Bureau of Statistics [ABS], 2006; Kaushal et al., 2007; Lee & Huh, 2004; Passero, 1996).

In Australian studies, housing has been found to place a major strain on the family budget as the family grows in size, although economies of scale have been observed with increasing numbers of children (Exter, 1992; Valenzuela, 1999). Like many other expenditure items, housing expenditure is influenced by employment status (Passero, 1996) and life-cycle stage (Carlucci & Zelli, 1998; Percival & Harding, 2005). However, certain characteristics of housing expenditure set it apart from other essential expenditures, with expenditure on this category differing markedly between households that are similar in every other aspect apart from tenure type (Carlucci & Zelli, 1998).

The studies that examine expenditure on furniture and appliances, insurance, telephone and the Internet are limited. Often these categories are grouped, or are not the subjects of the research focus, due to them comprising a small proportion of total household expenditure. Valenzuela (1999) observed economies of scale with increasing numbers of children for household furnishings, while a US study by Kaushal et al. (2007) found increases in spending on insurance associated with labour force participation; however, this finding may be consistent with US retirement savings policy, which requires working people to contribute to their own retirement pension. The authors' prior study (Brandrup & Mance, 2010) found little evidence of changed spending associated with the birth of newborns for these expenditure items.

In contrast, a prior Australian study has examined holiday spending in detail. Yusuf and Naseri (2005) found income, family life cycle (rather than family size) and ethnicity influenced holiday-taking behaviour. Families with preschool children were the most likely of all household types to take domestic holidays and the least likely to take overseas holidays, resulting in low overall combined expenditure on holidays. The authors did not find in their prior study (Brandrup & Mance, 2010) any significant evidence of changed holiday spending associated with the arrival of newborn children.

Similarly, in the authors' previous paper (Brandrup & Mance, 2010), there was no significant evidence of changed alcohol and tobacco spending associated with the arrival of newborn children, although other studies have found that the presence of children moves expenditure away from "adult goods" such as alcohol and tobacco. Other contextual factors that have been associated with decreases in alcohol and tobacco expenditure include: decreasing social opportunities due to increasing demands of child rearing, reductions in consumption for health reasons relating to pregnancy, marital status, and the number of children in the household (De Leire & Kalil, 2005; Valenzuela, 1999). When these items are separately examined, it appears that expenditure on tobacco products remains relatively constant across couple families with and without dependent children, while alcohol expenditure decreases markedly when dependent children are present (ABS, 2006).

Finally, as discussed previously, of particular interest in Brandrup and Mance (2010) and the current study is expenditure on electronic goods. This is because various media reports suggest that some families spend the government's baby bonus on electronic goods rather than on goods and services for the benefit of the newborn (see, for example, Haynes, 2008). In the authors' prior study, while no evidence of increased expenditure on electronic goods was observed for couples experiencing a first or second birth, weak evidence of increased expenditure on electronic goods was observed for couples experiencing the birth of a third or subsequent child. In the current study, the authors were able to examine this item again with the benefit of additional observations contributed by births occurring in the year prior to the 2008 HILDA survey.

The analytic framework

The strength of this study lies in the methodology applied and the longitudinal nature of information collected in the HILDA survey. Specifically, this study employs fixed effects regression models in its methodology. By comparing expenditure reports of the same couple families over multiple years, fixed effects regression models are able to exploit the longitudinal nature of the data to control for the characteristics of each household that have not been explicitly collected by survey instruments - providing those characteristics have not changed over the study period. These time-invariant characteristics may, for example, include some personality characteristics of household residents. However, some characteristics that might normally be thought of as time-invariant over a short time frame may in fact vary over the study period as a result of a major life event, such as the birth of a child, and this could result in changes in spending preferences. Such changes would not be controlled for in fixed effects regression models and would instead manifest themselves in the coefficients estimated by the models.

The study sample includes data from 2006 through to 2008 relating to female HILDA survey respondents, aged between 15 and 47 years in 2008, who lived in a couple household in all three years, without any other persons besides their own children, and were in a married or cohabiting heterosexual relationship with the same person in all three years.³ The selection process therefore excluded women who were not living with the same male partner⁴ in all three waves. Household data were incorporated into each respondent's record. This enabled the household to be the unit of analysis, because the selection process ensured that each respondent essentially lived in the same household throughout the period. Survey data prior to 2006 could not be used for this study because the HILDA survey spending categories were inconsistent with the categories used since that year (these spending categories are discussed later).⁵

After dropping cases with missing expenditure data or a multiple birth, the final sample consisted of 1,199 households, of which 349 included women who had given birth to or adopted 411 babies between 2006 and 2008. One hundred and seventy-four of the newborns had no older siblings in the household ("first-born"), 159 had one older sibling in the household ("second-born"). and 78 had more than one older sibling in the household ("third- or subsequent-born").

Two sets of fixed effects linear regression models were estimated to examine differences in expenditure associated with the arrival of a newborn:

• In the first set of models, spending reported in the survey in which a household experienced a birth was compared against spending reported by that same household in other years. For example, if a household in the 2007 survey reported that a newborn had arrived in the twelve months leading up to that survey, that household's spending in the 2007 survey was compared against the household's spending in the 2006 and the 2008 surveys. The effects of the birth of a first, second, and third or subsequent child were estimated separately. Strictly speaking, the first set of models do not estimate change in spending between years; the models instead estimate elevated or reduced levels of spending specific to the birth year. The difference in concepts is subtle; however, transitory changes in spending that occur in the birth year would be very specific to that year, while a sustained change in spending is less strongly specific to the birth year because the sustained spending is related more to the presence of a child rather than the birth event itself.
• In the second set of models, spending by households that reported in 2008 that a first-born child had arrived during the previous twelve months is compared against spending by the same households in the 2006 survey, with data from the 2007 survey being dropped from the models.⁶ Unlike the first set of models, the second set of models estimates changes in spending between two years. Although this approach poses methodological challenges, the second set of models allows for the removal of behavioural changes due to pregnancy, given that most mothers reporting a birth in the 2008 survey were pregnant when surveyed in 2007.

Dependent variables

The household spending section of the HILDA Self-Completion Questionnaires has, since 2006, listed 25 types of expenses on which Australians regularly spend their money. This study used the household spending items contained in the HILDA survey household file, which averaged the household spending responses if more than one person in a household provided responses (Watson, 2010). Respondents were asked to give their best estimates of the average amount that their households spent on 25 items - grouped into weekly, monthly and annual expenses. For annual expenses, respondents were asked to estimate how much the household spent on those items over the past 12 months. The 25 items were aggregated into thirteen spending categories, with each of those spending categories being the dependent variable in thirteen fixed effects linear regression models. Two other spending categories were created from household spending items found in the HILDA Household Questionnaire, thereby resulting in fifteen dependent variables in total. Summing these dependent variables does not equal total household expenditure because spending on many types of goods and services would not have been captured in the HILDA survey. The household spending responses in 2006 and 2007 were converted to 2008 dollars by adjusting for changes in the Consumer Price Index (CPI), using CPI figures broken down into the group of goods and services most closely corresponding to each expenditure item. Table 1 presents descriptive statistics on the fifteen dependent variables used in this study.

Notes: ^a Not annualised, as data on the number of typical weeks was not available.

Source: HILDA (2006-08), release 8.0

It can be seen from Table 1 that for most households the largest spending items were housing, groceries and transport. Spending on child care and education varied a great deal across households because most families reported spending nothing or small amounts on these services, while other families spent substantial amounts. Spending on furniture and appliances, electronic goods and holidays also varied considerably across households.

It is expected that the estimates of expenditure on these household spending items are subject to substantial measurement error, as they are self-reported and spending diaries are not kept (Gibson & Kim, 2007). For example, Headey (2008) found that the means of several 2005 HILDA survey expenditure items (clothing and footwear, health care, holidays, hobbies, and child care) differed by more than 10 per cent from the means of comparable items in the ABS 2003-04 Household Expenditure Survey (HES), which was based on spending diaries. More reassuringly, Headey also found that the other HILDA survey expenditure items, when totalled, differed by less than 4 per cent from the HES items after adjustment for inflation. As stated in the authors' prior study (Brandrup & Mance, 2010), model efficiency would be adversely affected by measurement error in the HILDA survey expenditure data. However, this only means that false negatives are more likely (i.e., the measurement errors are likely to lead to fewer statistically significant results), and any findings that some spending categories are significantly different in a birth year would remain valid. Nonetheless, some bias in the results is expected, and, when interpreting the study's results, more emphasis should be placed on whether some types of spending are significantly higher or lower in birth years, with less emphasis placed on the dollar estimates.

Independent variables

A range of time-varying independent variables were used to develop the models. The variables of interest in the study were: birth of first-born child, birth of second-born child, and birth of third- or subsequent-born child.⁷ These variables were used in preference to the option of using a single birth variable for the purpose of capturing birth order effects. All three birth variables were included in each model as it was important to control for the not-unusual situation of households reporting births in two consecutive surveys. Other control variables included: education,⁸ employment, income support and health status of women and their partners; household disposable income;⁹ geographical remoteness; and housing tenure.

Although not reported in detail here, descriptive statistics based on 2008 survey data differ to some extent from the 2007 survey descriptive statistics reported in the authors' previous study (Brandrup & Mance, 2010). Although both the 2007 and 2008 surveys generally indicate that households reporting the arrival of a third- or subsequent-born baby had a higher prevalence of characteristics normally associated with socio-economic disadvantage, none of those associations were statistically significant in the 2008 survey.¹⁰ Another difference emerged in relation to the growth in disposable incomes of households experiencing the arrival of newborn babies. The authors' previous study found that survey households experiencing the arrival of newborns in the 12 months prior to the 2007 survey had experienced higher disposable income growth between 2006 and 2007 than did other survey households, and had significantly higher disposable incomes than that of other households in 2007. However, the 2008 survey revealed a different picture - households that reported the arrival of a newborn child in the 12 months prior to the 2008 survey had disposable incomes that were not significantly different from that of other households in either 2007 or 2008.

Results

First set of models: All births over three years

The results for the first set of models are presented in Table 2. Each coefficient reported in the table is an estimate (in 2008 dollars) of how much spending in a birth year differed from spending in other years. The coefficients are not estimated levels of spending in the birth year - they are estimates of how much spending differs from "normal" spending in other years. A significantly positive coefficient of $100, for instance, would indicate that spending specific to the birth year was estimated to be $100 more than spending in other years. The results relating to control variables are not presented, but are available from the authors on request.

Notes: * p < .05, ** p < .01. a Not annualised, as data on the number of typical weeks was not available.

Source: HILDA (2006-08), release 8.0

Regardless of birth order, spending specific to the birth year was not significantly elevated or reduced for the categories of: groceries; meals eaten out and takeaway; education; furniture and appliances; general insurance; electronic goods; and telephone and Internet. There was a significant reduction in adults' clothing expenditure ($274 per annum) associated with the arrival of a second-born baby; and there was a significant elevation in children's clothing expenditure ($435 per annum) associated with the arrival of a first-born baby. There was a significant reduction in child care expenditure ($31 in typical weekly expenses) associated with the arrival of second-born babies.

Spending on health care was significantly elevated for each birth order, with the estimated elevation ranging between $532 and $641 per annum. Transport expenditure was at a high level with the arrival of first- and second-born babies, but these estimates were not significant. However, there was a significant elevation in transport spending ($4,922 per annum) associated with the arrival of third- and subsequent-born babies. A significant reduction in housing spending ($3,064 per annum) was associated with the arrival of first-born babies. There were also significant differences in holiday expenditure. The arrival of a first-born baby was associated with a reduction in holiday expenditure of $787 per annum, while the arrival of second-born babies was associated with a reduction in holiday spending of $470 per annum. A reduction in alcohol and cigarette spending associated with the arrival of first-born babies was observed, but this did not reach statistical significance.

For most spending categories, control variables had no effect on which coefficients were significant at the 95% level. However, the reduction in holiday spending associated with the arrival of second-born babies only became significant after controlling for the arrival of first-born babies in the preceding year. This implies that the reduction in holiday spending associated with the arrival of first-born babies tended to mask the reduction associated with the arrival of second-born babies if the arrival of newborns was reported in two surveys in a row. In relation to housing expenditure, the reduction in spending associated with the arrival of first-born babies only became significant once mothers' employment status was controlled for. Closer investigation revealed that this was because of a marginally significant (p = .051) positive association between spending on household renovations and women moving out of the work force.

Second set of models: First-born newborns in 2008

Table 3 shows changes in annual household expenditure (in 2008 dollars) associated with first-born babies by comparing 2006 spending (the pre-pregnancy year) with 2008 spending (up to twelve months after the birth event). As stated earlier, this allowed for comparison of expenditure prior to pregnancy with expenditure after the birth of a newborn, which eliminated changes in expenditure behaviour that might be due to pregnancy rather than the arrival of a newborn. Again, the results relating to control variables are not presented, but are available from the authors on request.

Notes: * p < .05, ** p < .01. 
^a Not annualised, as data on the number of typical weeks was not available.

Source: HILDA (2006, 2008), release 8.0

The birth of a first child was associated with an increase in grocery expenditure of $1,993 per annum, an increase in children's clothing expenditure of $875 per annum, and an increase in child care expenditure of $44 in a typical week. A large and significant increase in expenditure on health care associated with the birth of a first baby was also observed ($755). A significant decrease in housing spending ($5,315 per annum) was associated with the arrival of first-born babies. There were no significant changes in expenditure associated with the birth of a first-born child for any of the other spending categories.

For all but one of the spending categories, control variables had no effect on which coefficients were significant at the 95% level. The one exception was housing expenditure, in which the decrease in spending associated with the arrival of first-born babies only became significant once mothers' employment status was controlled for (as was the case in the first set of models).

Discussion

As mentioned earlier, due to study limitations, more emphasis should be placed on whether some types of spending are at significantly different levels in birth years, and less emphasis should be placed on the dollar estimates when interpreting the study's results.

Reports of actual expenditure by couple households between 2006 and 2008 indicate that families adjust their spending when a new baby is born. Spending on particular categories was found to vary with birth order, whether the item was discretionary or essential in nature, and was in response to expenditure demands and lifestyle changes associated with the arrival of a newborn.

Turning first to grocery expenditure, consistent with the authors' prior study (Brandrup & Mance, 2010), there were no elevations or reductions in spending observed in Table 2 specific to a birth year for any of the birth order groups. However, large increases were observed in Table 3 when comparing pre-pregnancy expenditure with post-birth expenditure for parents of first-born babies, suggesting that behavioural change had occurred during pregnancy. When considered alongside expenditure on eating out and takeaway food, it appears that formerly childless couples spent more on groceries in the pregnancy year and/or were consuming more meals at home in the year of pregnancy, and this level of expenditure continued into subsequent years.

While increases in grocery expenditure might not be expected in relation to second or subsequent newborn children, due to the low consumption of food by babies, the authors did expect baby-related "supermarket" products to increase grocery spending. It is possible that parents experiencing higher order births may have adjusted their expenditure to purchase lower cost grocery items to offset new baby-related expenses. However, it is also possible that parents did not include expenditure on nappies and formula when reporting expenditure on this item because the HILDA survey questionnaire instructs respondents to "include food, cleaning products, pet food and personal care products", but provides no guidance on whether to include other items that might be considered to be part of grocery spending. If respondents did not include spending on baby products such as disposable nappies and formula, or they used cloth nappies, grocery consumption would not be expected to increase.

The significant reduction in adult clothing expenditure in Table 2 specific to the birth year of a second child possibly reflects reduced shopping opportunities or couples shifting expenditure away from adult-focused goods. A reduction in spending was not observed for couples having first-born children, possibly because women may have been purchasing maternity clothes, replacing normal expenditure on adult clothes, during their first pregnancy. Some of these maternity clothes may have been used again for subsequent pregnancies. In contrast, consistent with the authors' prior study (Brandrup & Mance, 2010), children's clothing expenditure was significantly elevated with the birth of a first child, as formerly childless couples purchased essential clothing items for the new baby. There were no significant elevations in spending on children's clothing for couples having second or subsequent children, suggesting there was recycling of clothes for such children.

Table 2 indicates that there were no significant elevations or reductions in spending on child care specific to the birth year for a first child; however, in Table 3 there is a significant increase between 2006 and 2008 for families who reported the arrival of first-born children in 2008. Although the two estimates look as though they are in conflict, they are not directly comparable. As discussed earlier, the two sets of models are based on subtly different concepts. The first set of models (Table 2) estimated elevated or reduced spending specific to the birth year, and a sustained change in spending is less strongly specific to the birth year because sustained spending is related more to the presence of a child rather than the birth event itself. A close examination of the data showed that child care spending tended to be sustained, rather than transitory. Although this type of spending was zero for most families with a first-born newborn, a minority of families with a first-born newborn incurred child care expenses for the first time, and these expenses typically continued at a positive level for some years after the birth. However, the second set of models (Table 3) estimated change in spending between 2006 and 2008, and the sustained nature of child care spending had no effect on its significance. The results in Table 2 showing a reduction in child care expenditure specifically associated with the birth year for second-born newborns most likely reflect employed mothers taking up maternity leave or withdrawing from the workforce, and therefore having less need for paid child care when they have a second child.

Large and consistent elevations in health expenditure were observed in Table 2 for each additional birth, and are likely to reflect increased out-of-pocket health care expenses associated with pregnancy, birth and care of a newborn. When the pregnancy year is removed in Table 3, the expenditure change associated with pregnancy and childbirth is larger, implying that health care expenses must have already been at an elevated level during pregnancy.

Large but non-significant elevations in expenditure on transport for couples experiencing the birth of a third and subsequent child observed in the authors' prior study (Brandrup & Mance, 2010) became significant when an extra wave of data was included (Table 2); this is likely to be due to purchases of larger vehicles. Given the large average amount of transport expenditure increases observed for the other birth orders, a significant result was expected for all birth orders. However, the large variability in reports of expenditure indicates that transport costs vary greatly between individual households. For example, such expenditure may depend on whether older children attend school, the number of cars in the household, the availability of public transport, and the workforce participation of each parent (Kaushal et al., 2007; Passero, 1996).

There were also large average elevations and reductions observed for housing expenditure in Table 2, including a large and significant reduction in spending associated with the arrival of first-born newborns. A significant and even larger decrease associated with first-born newborns is reported in Table 3. Closer examination of the data reveals that reduced housing expenditure reported by parents of first-borns largely reflects formerly childless couples decreasing the amount they spend renovating their family home. However, as in previous studies, there is a large amount of variation in housing expenditure when individual study couples are examined (Carlucci & Zelli, 1998). The large amount of variation is the most likely explanation as to why changes in housing expenditure in the authors' prior study (Brandrup & Mance, 2010) failed to attain significance.

Similarly, in line with Brandrup and Mance (2010), no significant results were observed in Table 2 for expenditure on education, furniture and appliances, general insurance or telephone and Internet. In contrast, marginally significant results for expenditure on electronic goods became non-significant with increased sample size, dispelling media reports that couples spend their baby bonus on adult-focused consumer goods.

In Table 2, the amount of discretionary spending on holidays was at a significantly reduced level for couples experiencing the birth of a first- or second-born child, with childless couples reducing expenditure in the post-birth year compared with the pregnancy year. In line with expectations, spending may have been adjusted because of reduced opportunities for holidays due to the arrival of a newborn, or because of the need to purchase baby-related goods.

Finally, no significant results were observed for expenditure on alcohol and cigarettes in Table 2, although the large decrease observed in the skipped year model (Table 3) was - if disaggregated into separate alcohol and cigarettes spending - entirely due to a significant reduction in alcohol spending.

Conclusion

Information on the expenditure patterns of parents of newborns is important to a range of current policy debates, including those that seek to determine the required level of family assistance, influence women's attachment to the labour force and fertility rates, and the utility of and need for additional financial support around the birth of a child. Although expenditures in this study do not total to provide an estimate of the complete costs associated with the arrival of a newborn, these results do provide some indication of how parents in stable married or cohabiting relationships adjust their spending to meet the financial demands of pregnancy and newborns, and where these pressures may lie. For this large subset of families, this study shows that parents of first-born children report increased or elevated expenditure on groceries, health and children's clothing, but reduced levels of spending on holidays. Parents of second-born children report elevated expenditure on health, and reduced levels of spending on child care, adults' clothing and holidays. Parents of third- and subsequent-born children report elevated expenditure on health and transport. When considered together, these results indicate that parents adjust their discretionary spending to meet new demands for essential expenditures related to the birth of their newborn child. While this study does not comment on the merit of any particular policy it does provide evidence for the need for government support for parents around the birth of a child, particularly with respect to increased requirements for expenditure on health.

This study is restricted in its scope to stable couple families. Further research could examine whether lone parents and other parents not in a stable married or cohabiting relationship change their spending in different ways on the birth of their children.