The fiscal contribution of migrants

19 April, 2006

Summary
1. A Home Office report in 2002 claimed that migrants contributed 2.5 billion more in taxes in 1999/2000 than they received in benefits and State services.

2. This claim has been continually repeated to support the assertion that the government's policy of large-scale immigration is economically beneficial for the country.

3. The Home Office calculations were produced in a year in which the government's finances were in overall surplus so that both natives and migrants could be expected to make a positive fiscal contribution.

4. In 2005 the IPPR updated the Home Office's calculations for 1999/2000 and also examined subsequent tax years up to 2003/4. In doing so, they described the original Home Office estimate as "meaningless" because it had taken no account of the fiscal surplus in that year. For their update, the IPPR used substantially the same statistical basis as the Home Office had employed. To neutralise the effect of an overall budget surplus or deficit in a particular year, the IPPR measured the relative performance of the migrant and native populations. They found that in 1999/2000 migrants' fiscal contribution was 300 more per head than natives and that the difference had grown to 800 by 2003/4.

5. Both the Home Office and IPPR used calculations which apportioned dependent children to the migrant populations but only if both parents (or the sole parent) were foreign-born. Children who had 1 UK-born parent and 1 foreign-born parent were attributed entirely to the native population.

6. This meant that in "mixed" households of 1 UK-born and 1 foreign-born parent, the cost of all the children was attributed to the UK born parent and none of the costs fell to the foreign-born parent.

7. This clearly distorted the result. Migration Watch have therefore recalculated, apportioning the cost of children of "mixed" households equally to the native and migrant populations. The number of children involved is about 1.4 million, or 12% of all children, so the effect is substantial, indeed critical to the outcome.

8. On this basis the results were as follows:

9. We conclude that the Home Office study made a serious error in selecting the basis of their calculation. The government have since repeatedly claimed that migrants contribute 2.5 billion more to the Treasury than they cost. This claim, central to their case for massive levels of immigration is now disproved. A summary of the main findings is at paras 46 - 49.

The Home Office Study

10. In 2002 the Home Office carried out a study into the fiscal contribution of migrants in the UK [1]. The fiscal contribution is the amount which migrants contribute in taxes and other revenue compared to the value of the benefits and state services which they receive. The migrant population was defined as the foreign-born population plus their dependent children (those aged 0-15). The study described its findings as 'necessarily tentative' because of major information gaps.

11. In summary, the findings were that in 1999/2000 migrants in the UK contributed 31.2 billion in taxes and consumed 28.8 billion in benefits and State services, thus producing a net fiscal contribution of approximately 2.5 billion after rounding. This was described as being equivalent to about 1p on the basic rate of income tax.

12. A number of criticisms of the Home Office report were subsequently made. The most important of these was that the Home Office had used a financial year (1999/2000) when the public accounts were in surplus so that both the migrant population and the UK-born population would have made a net positive fiscal contribution in that year.

13. This criticism was subsequently addressed in work carried out by the Institute of Public Policy Research (IPPR).

The IPPR Study

14. IPPR carried out a further analysis of the fiscal contribution of migrants in 2005 [2]. They were supplied with the data used by the Home Office in their study and they used the methodology in that study as the basis for their work but with a few technical changes. They updated the 1999/2000 calculations using updated data [3] and they extended the calculations to the next four tax years (2000/2001 to 2003/4 inclusive).

15. To avoid the criticism that the Home Office had chosen a year when the public finances were in surplus, IPPR used as their main measure the ratio of receipts from migrants to expenditure on benefits and services for migrants. They compared this with the ratio of receipts from the UK-born to expenditure on benefits and services for the UK-born. They labelled this ratio the Net Annual Fiscal Index (NAFI).

16. Their findings using this measure were:


17. Thus in 1999/2000 their calculations showed that migrants contributed 106 in taxes and other revenue for every 100 of benefits and State services they received. This was a superior performance to the UK-born who, in that year, contributed 101 for every 100 of benefits they received.

18. By 2003/4 the public finances had gone into an overall deficit and both migrants and the UK-born population were contributing less than they were receiving in benefits and services. However, according to the IPPR, the migrant population was contributing 99 for every 100 received compared to just 88 for the UK-born population. The relative contribution of migrants had therefore improved from being about 5% better than the UK born in 1999/2000 to about 11% better than the UK-born in 2003/4.

The Home Office and IPPR methodology

19. The Home Office and IPPR have used essentially the same methodology in their respective analyses. They have each taken Treasury figures on expenditure and receipts and split this between the migrant and UK-born populations, using data from the Labour Force Survey or by population shares. For example health care costs have been split by population shares by age groups using Department of Health figures on the unit costs of health care for different age groups. Similarly primary and secondary education costs have been split using the population shares in the 5 to 15 age group. Benefits such as unemployment and housing were split according to the percentages of migrants and UK-born receiving those benefits (from the Labour Force Survey). More general costs, such as defence costs, were split using the overall percentage of the population who are migrants. Tax and National Insurance Receipts were directly computed from detailed data on the pay of migrants and the UK-born from the Labour Force Survey and grossed up to reflect the whole population.

20. Full details of the basis which the Home Office used are available in Appendix A of their report. Some small changes were made to this methodology by the IPPR; these are detailed in pages 3 to 6 of their report. (The principal change was to exclude corporation tax receipts attributable to foreign ownership of UK quoted companies from the corporation tax revenue figure before apportioning it between migrants and the UK-born).

21. The use of the Labour Force Survey is therefore critical to both these studies. The Labour Force Survey provides very valuable data and is used widely by government and research organisations for statistical and social analysis. It is, however, based on a voluntary survey of households. In the context of migration the voluntary nature of the survey is likely to exclude those who are in the UK illegally (eg. through clandestine entry or as visa over-stayers) and those who are seeking asylum or have had their asylum claim rejected. The often repeated suggestion that the outcome for migrants is reduced because the migrant population includes asylum seekers who are not permitted to work is not, therefore, applicable to this study.

22. Clearly, the definition of a migrant is critical. Both studies have defined a migrant as a person born outside the United Kingdom. There is considerable discussion in both reports about how to treat dependent children (0 to 15 year olds) of migrant parents. They both come to the conclusion that UK-born dependent children should be included in the migrant population because: a) they are a considerable cost and b) they would not have been here if their parents had not migrated to the UK. However, they have attributed UK-born children to the migrant population only where there are two adults on which the child depends and both were foreign-born or where there is a single foreign-born adult on which the child depends. Children living in "mixed" households where one adult is foreign-born and one is UK-born have been attributed entirely to the UK-born population.

23. It should be noted that this basis was specified clearly in the Home Office RDS Paper 77 (albeit in the Appendix). However, it was not made clear in the notes on the tables sent to the IPPR by the Home Office; these implied that children of "mixed" households had been apportioned between the migrant and UK-born populations on a 50:50 basis.

24. Perhaps as a result of this ambiguity in the letter they received from the Home Office, the IPPR have used the same basis as the Home Office. Yet the basis for allocating dependent children is critically important to the outcome as described in the next section.

Basis for allocating children of UK-born and Migrant Parents

25. Dependent children are particularly important to these calculations because they are a considerable cost (principally in respect of education) but do not, of course, contribute to direct taxation receipts (income tax and national insurance).

26. Data on children obtained from the Labour Force Survey shows there are nearly 1 million children from "mixed" households who have one parent who is UK-born and one who is foreign-born, as shown below [4]:



The expenditure on primary and secondary education in 1999/2000 was 23.625 billion [5]. So, just allocating children from "mixed" households on a 50:50 basis between the UK-born and migrant populations would add over 1 billion to the costs attributable to migrants [6] - wiping out over 40% of the surplus (of revenue over expenditure) attributed to migrants in the Home Office study.

27. The effect of allocating all children of "mixed" households to the host community is that the UK born parent is regarded as responsible for all the couples' children while the foreign born is responsible only for him or herself. The impact of this can be illustrated by considering a "mixed" household with two UK-born children. Suppose both parents are earning identical amounts and paying equal amounts of taxation and national insurance of, say, 8,000 each. Let us further assume that each parent receives State benefits and services of 3,000 and that each child receives services of 5,000 (principally in the form of education). By the Home Office/IPPR calculations the UK-born parent would be making a net fiscal contribution of -5,000 (their revenue contribution of 8,000 less the services they personally receive of 3,000 less the services received by the children of 5,000 each). The foreign-born parent would, in contrast, be making a net contribution of +5,000 (their revenue contribution of 8,000 less services they personally receive of 3,000).

28. This result, quite clearly a nonsense, illustrates the distorting effect of the method adopted both by the original Home Office study and the subsequent IPPR report.

29. Migration Watch has therefore re-calculated the fiscal contributions of the UK-born and foreign-born based on dependent children of "mixed" households being split 50:50 between the UK-born and foreign-born populations. (Migration Watch's basis also implicitly allocates foreign-born children of UK-born parents to the UK-born population - a logical corollary of allocating UK-born children of foreign-born parents to the foreign-born population. Our understanding of the IPPR/Home Office calculations is that foreign-born children are allocated to the foreign-born population irrespective of the birthplace of their parents). The results are outlined below:

Migration Watch calculations

30. Migration Watch has carried out three calculations - a direct comparison with the Home Office's calculation for tax year 1999/2000 and comparisons with the IPPR calculations for 1999/2000 and 2003/4.
a) Comparison with Home Office RDS paper 77 calculations

31. In the first we based our calculations on the method specified by the Home Office in RDS Paper 77. We requested the detailed methodology, cost data and calculations under a freedom of information request. All of this data is now publicly available on the Home Office website [7]. The Home Office advised us that the figures used in the spreadsheet '.. are now out of date; the Labour Force Survey time series was re-weighted in mid-2004, to take account of the 2001 census. We have been asked by ONS to recommend strongly that any future work you carry out should use the updated LFS values.'

32. We duly followed this advice and obtained fresh information from the Labour Force Survey for 1999/2000. However, we used the same Home Office data on costs and we also used the data which they had obtained from sources other than the LFS (such as University tuition fees). We thus ensured that the calculations would follow those of the Home Office as closely as possible. We did, however, allocate children of "mixed" households 50:50 to the UK-born and foreign-born populations [8].

33. Our calculations showed that these changes would alter the net fiscal contribution of migrants from the original Home Office estimate of a 2.5 billion excess of revenue over expenditure to a 200 million deficit.
b) Comparison with IPPR 1999/2000 calculations

34. The IPPR calculations for 1999/2000 differed slightly from those of the Home Office in so far as they used updated Treasury expenditure and revenue data and they apportioned corporation tax receipts between the two populations after reducing these receipts by the amount of corporation tax attributable to foreign ownership of shares in UK companies. This gave a reduced fiscal contribution for both the native and UK-born populations. Migration Watch have carried out the same calculations but allocating children of "mixed" UK born / foreign-born households in three different ways.

- all to the host community (as in the H.O. and IPPR studies)
- 50:50 between the host and migrant communities
- all to the migrant community.

The results are summarised on the following page:


35. The right hand column is included to illustrate the very large impact of the allocation of children of "mixed" households. . We think it would be illogical and misleading to allocate all these children to the migrant community. Equally, we think it illogical and misleading of the Home Office and IPPR to have allocated them all to the host community.

36. The middle of the three data columns which allocates these children on a 50:50 basis is, in our view, a much more accurate reflection of the fiscal contribution of the migrant population. It shows that:
c) Comparison with IPPR 2003/4 calculations

37. Migration Watch calculated the fiscal contribution of migrants for the financial year 2003/4 using the same basis as that used by the IPPR but then also with children of "mixed" households" allocated 50:50 to the UK-born and foreign-born populations. Our results are compared with IPPR's in the table below:



38. The Migration Watch result shows that the net fiscal contribution of the migrant population improved relative to the native population between 1999/2000 and 2003/4 (line 3) but only to the extent that it is now very similar to the native population (line 6) - not 12 % superior to it as IPPR had purported to show.

Comment on improvement in migrants' relative fiscal performance between 1999/2000 and 2003/4

39. Migration Watch calculations confirm one of IPPR's findings - namely that there has been an improvement in migrants' fiscal contribution relative to the UK-born between 1999/2000 and 2003/4. However, the improvement is relatively modest and it should be treated with some caution. Net migration into the UK rose dramatically from 1998 as the following graph demonstrates [9]:



40. Much of this increase was as the result of deliberate government action to stimulate immigration. For example, there was a steep rise in the number of work permits issued as shown below [10]:



41. Net immigration into this country adds predominantly to the age groups 15-24 and 25-44 as the following table for 2002 immigrants shows [11]:


42. The immediate fiscal effect of migrants of working age entering the UK is likely to be positive. Most of these migrants will be economically active and they will be making a fiscal contribution but, as they have very few dependants, their demands on government benefits and services will be low. However, this situation will change rapidly as they have families of their own and as their entitlement to benefits, and eventually pensions, takes effect.

43. Another measure of this effect is to look at the ratio of migrants of working age (16-64) to dependent children of migrants. In 1999 there were 234 migrants of working age for every 100 children but in 2003 this had increased to 278 for every 100 children [12].

44. In a situation where the net inflow of migrants is relatively stable the increasing demands of the earlier migrants will be offset by the reduced demands of the new migrants. However, when there has been a step change in migration levels, as there has been in recent years, there will be a temporary fiscal benefit. But it is one that can only be maintained by further increases in net migration levels. Such changes would be unsustainable.

45. The fiscal contribution of migrants as calculated for the financial year 1999/2000 followed a long period of consistent levels of net migration is therefore likely to be a better long-term indicator than the 2003/4 results which followed a sharp increase in immigration.

Summary of findings

46. The Home Office and IPPR studies of the fiscal contribution of migrants made a serious error in selecting the basis for their calculation. This crucially affected the outcome.

47. Adjusting it to the more logical basis that dependent children of a migrant and a native parent are allocated 50:50 to each population reverses the findings of the original Home Office paper. That is, far from making a net positive fiscal contribution of 2.5 billion in 1999/2000 as the Home Office claimed, the net fiscal contribution was negative at minus 200 million.

48. The IPPR findings, that migrants make a larger fiscal contribution than natives, are also shown to be invalid. In 1999/2000 their contribution per head was about 300 less than that of the native population not approximately 300 more as IPPR had suggested. By 2003/4 their contribution per head had improved but only to a level which matched that of the native population (instead of being 818 more as IPPR had calculated.)

49. There was an improvement in migrants' relative fiscal performance in relation to the native population between 1999/2000 and 2003/4 but this was likely to be as a result of the step change in the levels of net migration to the UK which has occurred since 1997. The improvement is likely, therefore, to be transitory in nature.

Conclusion

50. The Government has repeatedly claimed, in support of their policy of encouraging large-scale immigration, that migrants make a positive fiscal contribution of 2.5 bn a year. This report shows that central claim to be false.

---