Voluntary Evacuation on the Outbreak of War
England and Wales

(Chapter VII)

Under the official plans, 1,298,325 children, mothers and certain special groups were evacuated from the vulnerable areas of England and the outbreak of war.

The number of individuals who made their own arrangements to go to private houses, hotels and boarding houses in the safer areas of the country, and who evacuated themselves from London and other areas before the war and during the first week or so of September, was very large. This problem of private evacuation, in trenching on the supply of billets for mothers and children under the official scheme, was, throughout the history of evacuation, a continual source of worry to the Government. It is not, however, possible to make an accurate estimate of the amount of private evacuation after the year 1939, owing to the disturbances to the statistics created by extensive population changes, enlistments and other factors. This appendix is therefore devoted to a study of the movement between midsummer 1939 and the date of national registration, namely 29th September 1939. The result, when compared with the volume of official evacuation, affords, however, some guide to the quantity and direction of private evacuation at other periods during the war.

Six of the larger evacuation areas in England were first selected for analysis. The loss of population between mid-1939 and the end of September, after making allowances for enlistments1 and natural increase, was:

Table 1

Loss of population
Greater London
Liverpool and Bootle county boroughs
Birmingham and Smethwick county boroughs
Manchester and Salford county boroughs
Leeds county borough
Sheffield county borough


This figure includes of course both official and non-official evacuees. It is considerably less than the number of people who actually left these areas


owing to the return movement between 3rd and 29th September. The size of this return movement has now been estimated—first for the official evacuees.

When a national evacuation count was taken in the reception areas on 8th January 1940, it was found that forty-four percent of the unaccompanied children in England and Wales had returned. Assuming the return movement was equally distributed over the 127 days then twenty percent of the returning children would have left the reception areas by 29th September. By 8th January 1940, eighty-eight percent of the mothers, eighty-six percent of the accompanied children, eighty-one percent of the other classes, and fifty-five percent of the teachers and helpers had returned.

An earlier estimate of the return movement was made by the sending authorities on 5th December 1939.; This showed that for each of the six areas in question the proportion returning by 5th December 1939 was, for the three important classes:

Table 2

Proportion returning home

Unaccompanied children
Accompanied children

Greater London
Liverpool and Bootle
Birmingham and Smethwick
Manchester and Salford

For the three classes as a whole and for all six areas, the rate of return worked out at forty-three percent or, broadly, fourteen percent per month. By January 1940, when the national count was taken, it was found that fifty-nine percent of all the evacuated classes in all areas of the country had returned. As the areas included in Table 2 account for the majority of evacuees, and can therefore be accepted as representative of all areas, it may there be assumed that during he thirty-four days between 5th December 1939 and 8th January 1940 a further sixteen percent of the evacuees returned. This, however, was mainly because of a much greater rate of return among mothers and accompanied children during December and the Christmas period. Between the two days, 5th December 1939 and 8th January 1940, the proportion of unaccompanied children returning rose from thirty-three percent to forty-four percent among mothers from fifty-eight percent to eighty-eight percent,3 and among accompanied children from fifty-seven percent to eighty-six percent.3 Apparently, Christmas was a very important influence in determining the rate of return among mothers with young children.


It is clear that the rate of return, among the different groups and back to different areas, was not evenly distributed over the first three to four months of the war. Among unaccompanied children the rate appears to have been heavy during the first few weeks;4 the flow gradually declined and it does not appear to have been significantly affected by the Christmas period. Perhaps this was because some of the children went home for Christmas, returning after the holidays to their foster-parents. Among mothers with children there occurred an immediate and heavy return in September. The drift back subsided October and November but rose considerably in December, so much that only thirteen percent remained in the reception areas on 8th January 1940.

On the evidence presented here and from a study of many reports from reception areas it has been assumed that, of the return to all areas by 5th December 1939, forty percent of the returning unaccompanied children left the reception areas by national registration day and sixty-five percent of the mothers and accompanied children. In actual numbers, this assumption means that of 738,770 unaccompanied children sent on the outbreak of war 98,500 had returned by 29th September, the corresponding figures for mothers and accompanied children being 408,930 and 154,167. The combined percentage return by 29th September is therefore assumed to have been twenty-two percent. This figure probably errs on the low side, particularly if the experience of Cambridge, Glasgow and other areas that kept careful records was representative. The true figure may have been nearer forty percent.

Proceeding, however, on the assumption that of all those evacuated  twenty-two percent had returned by 29th September, calculations were then made of the number of official evacuees, in each class and for each of the six areas, who were still away from their homes on 29th September 1939. The difference between the figures thereby reached and the loss of population given in Table 1 represented (after allowance had been made for the small number of other classes officially evacuated) the number of private evacuees who had not returned by 29th September 1939.

The next step was to estimate the drift back among private evacuees during September. In the absence of any statistics, it has been assumed that private evacuation was composed of mothers and children in the same proportions for the different areas and for the total movement as for official evacuation. This of course was not the case, as national registration showed that the additional population in many reception areas on 29th September included a considerable number of adult men and elderly women. It is therefore arguable that the return movement in September was higher among private than official evacuees. The heavy weighting of official evacuation with unaccompanied children and their relatively


slower rate of return is the basis for this argument. Nevertheless, for the purpose of this analysis some figure had to be adopted, and it was therefore assumed that the rate of return was proportionately the same in both groups.

The number of private evacuees from the six major evacuation areas still remaining in reception areas on 29th September 1939 having been calculated, the figure was then increased by the estimated volume of return during September. A total of 1,311,300 (1,200,000 from Greater London) was thus reached. This figure was then stepped up in the same areas bears to those officially evacuated from the six areas bears to those officially evacuated from all evacuation areas in England on the outbreak of war. A total of 1.808.300 was thus obtained.

This figure of 1,808,300 private evacuees excludes the not inconsiderable movement from inner London to neutral areas in Greater London;5 it excludes Scottish movements, and it rests on certain favourable assumptions concerning the rate of return during September.

To check this figure the problem was investigated from the reception end, namely, the increase of population by 29th September, as disclosed by national registration.

  1. For all reception areas in every receiving country in England and Wales the difference in population was calculated for the period mid-1939 to 29th September 1939.
  2. Allowance was made for natural increase and enlistments.
  3. To the number of official evacuees known to have been received by each country at the beginning of September, the September drift back assumptions were applied and the resulting sum was deducted from the calculated additional population.l
  4. The balances for each county were taken to represent the number of private evacuees still away on 29th September 1939. The results were then multiplied for each county by the factor 100/78 e.g. the assumption was made that the drift back among private evacuees was the same as for official evacuees, namely twenty-two percent.

The result of this arithmetic was a total figure of 1,514,500 private evacuees. This figure, reached by estimating the inward flow, is lower than that (1,808,300) arrived at by estimating the outward flow. While the figure of 1,514,500 appears to understate the actual volume of private evacuation, it does confirm that the total movement was large and that it ranged between 1,500,000 and 2,000,000. The rate of return during September among private evacuees was in all probability higher than that for the


officially evacuated (twenty-two percent) and may have reached forty percent. If a percentage return of forty is applied then the figure rises from 1,514,500 to 1,969,000. It is also important to note that the calculation of the inward flow excludes all movements from evacuation to neutral areas, such as from London to neutral areas in parts of Essex, Hertfordshire, Kent, Middlesex and Surrey, and similarly in the provinces, namely, neutral areas in Cheshire, Derbyshire, Durham, Lancashire, Northumberland, Nottinghamshire, Southampton, Staffordshire, Warwickshire, Worcestershire, Yorkshire East Riding, Yorkshire North Riding, Yorkshire West Riding, Glamorganshire and Monmouthshire. If the effect of this movement into all neutral areas could be assessed it would add considerably to the figure of 1,514,500 (or 1,969,000).

What cannot be allowed for in this study is the number of people who were on holiday in reception areas on 29th September and who might (depending on the address given to the national registration officials) be counted in this analysis as 'private evacuees'. It is unlikely, however, that the number of such holidaymakers could have been sufficiently large after four weeks of war to significantly affect the broad conclusions drawn here.

From this examination of the available statistics it can reasonably be stated that, in addition to the 1,330,000 persons officially evacuated in England, nearly 2,000,000 persons moved under private arrangements. The figures for some counties are particularly interesting. The additional population in the reception areas of Devonshire on 29th September 1939, was 64,556 (After making the appropriate adjustments). The number of official evacuees sent to that county was 10,440. Thus, if no private or official evacuees had left the county by 29th September 1939, private evacuation outnumbered the official movement by 5 to 1.But if allowance is made for some return during September then roughly 71,800 private evacuees6 went to Devonshire as compared with 10,440 official evacuees. The ratio of private evacuation to official evacuation was also strikingly high in such counties as Cornwall, Somersetshire, Gloucestershire, Herefordshire, Buckinghamshire and Sussex East and West.7 This agrees substantially with the results of the analysis of the geographical distribution of reserved accommodation revealed by the February 1939 survey when 1,100,000 people had reserved rooms.8 Both investigations show that private evacuation to the western half of the country was much great than that to the eastern half. The figures for the reception areas of Wales are also interesting. Approximately 56,000 official evacuees were received in Wales in September. The additional population on 29th September 1939 was shown to be 132,000. If adjustments are made for a return movement in September then at least 120,000 private evacuees went to Wales.9 This figure may be compared with an estimate made for


the Committee of Imperial Defence in 1938, which concluded from railway statistics that at least 150,000 people arrived in Wales at the time of the Munich crisis.

Note on the assumption of a twenty-two percent drift back of official evacuees during September 1939.

Reasons have been given for thinking that a percentage of twenty-two errs on the low side. This view is supported by a study of the payment of billeting allowances and the recovery of such allowances during the period September 1939 to February 1940.10

(A) The Payment of Allowances. The average weekly payments for the billeting of all evacuated persons were approximately:




Of the total fall in average weekly payments between September and December, fifty-nine percent applies to the October figure. No data are available showing the division of these sums between unaccompanied children and mothers with accompanied children, but there is no doubt that the bulk of the cost was accounted for by the payments made for unaccompanied children.

(B) The Recovery of Allowances.

For the four-week period ending
Number of schoolchildren in respect of whom payments were made or who were on 'Nil Assessments'
25th November 1939
23rd December 1939
20th January 1940
17th February 1940

The recovery scheme began to operate in the last week of October and some delay in the complicated work of recovery and assessment explains the rise in the December figure. It is impossible to say what part of the December figure should be transferred to November to correct this distortion. These figures should, however, be compared with the total number of unaccompanied children evacuated at the beginning of September, namely 738,770. Unless the machinery for recovering allowances from parents was grossly at fault, and large numbers escaped payment and assessment, the magnitude of the difference between the number of children actually evacuated and the number of recoveries and assessments strongly suggests that the bulk of the draft back occurred in September and early October 1939.



Both these inquiries point to a greater percentage return by 29th September than twenty-two. The higher this return is fixed the greater will be the volume of private evacuation. If, for instance, a figure of thirty percent is adopted for official evacuees then the forty percent return for private evacuees suggested above appears to reasonable. In that case, the amount of private evacuation at the beginning of the war can be put in round figures at 2,000,000 persons for England and Wales


Contents  *  Previous Chapter (Appendix 1)  *  Next Chapter (Appendix 3)


1 National registration excluded non-civilians and the population of ships in or nearing port. These excluded sections are estimated to amount to 2.2 percent of the total population for the whole of the United Kingdom—see National Register: Statistics of Population, 29th September 1939 (1944).

2 These figures look suspiciously like guesswork. The numbers involved, however, are so small that any error would not significantly affect the conclusion drawn from this table.

3 It was estimated in the Ministry of Health that by 24th October 1939 about forty-five percent of these groups had returned.

4 Cambridge, for example, received 6,700 evacuees. In September 2,500 returned and in October 800. The monthly rate of return thereafter continued to fall. In Glasgow, forty-three percent of the evacuated schoolchildren had returned by 5th November. The corresponding proportion for all Scottish areas was thirty-eight percent. The rate of return among mothers and other children was much higher, for by 5th November, sixty-three percent of them had returned to the Scottish evacuation areas. For other evidence of a heavy return in September see pp. 5489.

5 Neutral areas surrounding London showing a reduced population on 29th September may be estimated to have at least 84,000 people, while other neutral areas in the same region had an aggregate rise of about 69,000. Thus, without allowing for any return movement, it may be presumed that about 150,000 people moved into or out of the neutral areas surrounding London.

6 On the assumption of a twenty-two percent drift back and ignoring private evacuation into neutral areas of Devonshire.

7 Conversely, the ration was low in Lancashire, Durham. Cheshire, Suffolk East, Bedfordshire, Cambridgeshire, Huntingdonshire, Nottinghamshire and Leicestershire.

8 See Chapter III, pp. 37-8.

9 On the assumption of a twenty-two percent drift back, and ignoring private evacuation into neutral areas of Wales.

10 For England and Wales only.

Transcribed and formatted for HTML by David Newton, HyperWar Foundation