Working Paper 436: ‘The Costs of Working in Ireland’ (available at Sussex University) has caused considerable controversy. The part of the conclusion that has drawn the most attention is this:
This is important for policy as it substantially affects the incentives to work. A comparison of take home pay shows that it does not pay to work for 1% of the population. However, a comparison of take home pay plus extra expenditures shows that 15% of the people without children, and 44% of people with children, are better off not working.
The paper aims to calculate the additional costs incurred when working, and thereby calculate the incentive or disincentive to taking a job instead of staying on Jobseeker’s Allowance. It has since been withdrawn by the ESRI.
It uses data from the Household Budget Survey 2004-2005. This data is used to estimate what level of additional costs are associated with working. This dataset was probably chosen because the latest Household Budget Survey is not available in final form (Preliminary results for 2009/10 available here). A repeat of the study with 2009-10 might yield different results (for example, transport expenditure has dropped 5%, while heat and lights expenditure has increased by 15%).
The study also uses SILC data. The paper does not appear to state which SILC data was used, so it is likely to be data from 2004 or 2005, to match the HBS. This data gives information on income for a sample of people (a different sample from the HBS). This income data is used to estimate what each person would earn if employed net of additional costs (as obtained previously), and estimate what each would earn if unemployed. The two are compared to obtain the reported 44% of parents who would be better off on the dole. The paper says that this is illustrative only, since the two datasets relate to different sets of people.
When we look at the description of the population in the SILC dataset (2005, table 8, 2004 figs) we see this:
Assuming we take for analysis all people over 16 who are employed (40.8%), unemployed (3.3%) and home-makers (14%), then the vast majority of people for whom this comparison was done were in employment. Those unemployed were a small minority.
Therefore reports such as this in the The Sunday Business Post website are misleading:
The paper claims that 44 per cent of parents in Ireland with young children under 5 are better off staying on the dole then going back to work.
Since the SILC population were mostly working, and only a minority of the non-working were on benefits, it seems that the disincentive effect found in the paper did not in fact disincentivise working. The people involved could not “stay” on the dole – they were not on the dole.
A final comment – the details of this study cannot be assessed without redoing the calculations involved. The entire HBS dataset was not used, and classification of expenses was done based on employment status, which is not data available in the reports from the CSO. However there is at least one question that arises from browsing through the CSO HBS 2004-5 full tables report.
Table 2 gives household information and expenditure by income decile – in other words the households surveyed are split into ten groups, based on income. Here is the spread of children between households in these income groups:
Here is the distribution of childcare and nursery/montessori costs.
Looking at the 9th and 10th decile, childcare costs are much higher in the 10th decile, even though the 10th decile has more people doing “house duties”, and a very similar number of children. This suggests that childcare costs are affected by income – higher income allows selection of more expensive childcare. The overall shape of costs is steeply upward.
This raises the question – what is the marginal cost of employment? Is everyone using the same type of childcare solutions? The Quarterly Household Survey, Q4 2007 on Childcare has some interesting insights here. In couples where both worked fulltime, 42% with children under 12 managed with no non-parental childcare at all (this included 32% of parents of pre-school children). Where one was working part-time use of childcare dropped further. In addition, unpaid relatives provide unpaid childcare in some 10% of households.
This does not mean that the Irish childcare system is acceptable – 60% of households disagreed with the statement ‘I have access to high quality, affordable childcare in my community’. But it does mean that a variety of approaches are used by households to care for children, and the marginal cost for a individual can be low in financial terms, though high in human costs.
Even where childcare costs make working prohibitive this does not necessarily lead to increased numbers on the dole. The 7th decile and the 9th decile have similar numbers of children per household, yet the costs for childcare are different (the ninth are almost twice as much). In this instance there are nearly twice as many people on “home duties” in the 7th decile per household. So it is plausible that childcare costs are a disincentive to work for these people, but instead of going on benefits, the people (mostly women) are leaving the work force altogether. (Of course, the causation may work the other way – they chose not to work, and this reduces the household income. This data cannot tell us).
In similar ways, average transport costs or average use of takeaway food may not reflect the marginal cost of employment for a particular individual, where cheaper options may be substituted. Neither does it mean that people who stay out of paid employment will go on the dole. This report high-lights issues, particularly around childcare and transport, where tough choices may need to be made by working people. However given the data was taken at a time when employment was very high, it is clear people will work, even when it “does not pay”. Use of this report to explain the current levels of unemployment involves making unjustifiable inferences.