Joint Core Strategy - Proposed Main Modifications

• In all three projections the rate of growth in the number of households falls through the projection period.
• The extended 2011 DCLG projection and the two scenarios all suggest that, given both the demographic and economic changes that have taken place, the 2008 projections lie above likely expectations. On the other hand the 2011 extended projection would appear to be pessimistic in that it does not allow for any catch up in the future in comparison to past long term trends.
• The ‘partial return to trend’ and the ‘2008 tracking’ approaches give fairly similar results with estimated annual growth in household numbers that is some 10% below the 2008-based projection.
• It is suggested that this would be a prudent planning assumption. However, what actually happens should be monitored closely and be re-assessed at each review point during the plan period

3.7.   Note that:

• a. The 2011 census found 9,900 or 6.0% fewer households than the 2008-based projections had suggested. This is a much larger percentage difference than the population difference, implying that household formation rates (which measure the tendency of groups of people to form households) were significantly lower.
• b. The projected growth in the number of households in the area between 2011 and 2021 in the 2011-based projections is significantly lower than in the 2008-based projections – by 19.4%.

3.8.   In working with the 2011-based population and household projections it is important to recognise that they are interim projections. The 2011-based population projections take as their starting point population numbers and age profiles for local authorities from the 2011 census. They then apply the trends in fertility, mortality and internal migration estimated for the 2010-based population projections: they were not based on a full re-working of the trends. This was done to enable projections reflecting the census to be produced relatively quickly and before the census data to update the trends were available. Adjustments were also made to estimates of international ‘out’ migration at the local authority level as a result of an improved methodology for estimating which local authorities those leaving the country had come from. However, the total number of ‘out’ migrants for England as a whole remained unchanged.

3.9.   The ONS make it clear in the statistical release that accompanied the 2011-based projections that caution should be exercised in using certain aspects of their projections. The main concern referred to was the projection of birth rates, which will not significantly affect the number of households before 2031. There may, however, be a similar issue affecting the projection of flows from one authority to another in the UK – a factor that can have substantial consequences for housing requirements. The note on this states:

“Differences in the age structure at local authority level have also resulted in changes to projected levels of internal migration, that is, people moving their area of residence from one local authority to another within England. This is because migration rates based on historic trend data are applied to the new population base. Where the size and structure of the new population base in a local authority is very different from the 2010-based projections for 2011, particularly at ages most likely to migrate, the applied migration rate may over or underestimate the number of people moving from an area.”[6]

3.10. The net internal migration flow into or out of a local authority is often a relatively small difference between two much larger gross ‘in’ and ‘out’ flows. A small percentage change in either the projected ‘in’ or ‘out’ flow can therefore result in a large change in the projected net flow, with sizeable consequences for the projected change in population.

3.11.   Charts showing how the projections for births, deaths and migration flows have changed between the three population projections are included in Appendices 2-4 which give more detailed data for the three authorities.

3.12.   It should also be noted that in the statistical release for the interim projections DCLG state that, “The household projections should be used as part of the evidence base regarding the future demand for housing….”, the inference being that they should not be used in isolation, but alongside other evidence.

3.13.   In addition it should be noted that both ONS and DCLG are clear that their figures are projections and not forecasts: i.e. the estimates indicate what would happen if past trends were to continue; they are not stating a view on what is most likely to happen. For example, in the statistical release on the 2010-based Sub-national Population Projections, ONS state that the projections are, “…not forecasts and do not attempt to predict the impact that future government or local policies, changing economic circumstances or other factors might have on demographic behaviour”[7] .

Extending the 2011-based headship projections

4.7.   The household formation rates which were published with the 2011-based household projections have been extrapolated using simple statistical techniques. The chart below shows the result for the overall household formation rate for West Northamptonshire, with the 2008-based rates shown for comparison.

4.8.   The extrapolated 2011-based headship rates show much slower growth in the overall household formation rate compared with long term trend since 1991.

Extending the 2011-based interim household projections

4.9.   Applying the extrapolated headship rates to the extended 2011-based population projections produces the extended 2011-based household projections, shown in the chart and table below.

4.10.   The extended 2011-based household projections show a similar picture to that in the official figures for 2011-21, albeit with a slightly larger difference between the projected household growth numbers. Over the period 2011 to 2031 the 2011-extended projection suggests household growth that is 20.7% slower than the 2008-based projection (compared with 19.4% over the period 2011 to 2021).

4.11.   It should be noted that in both the 2008-based projections and the extended 2011 projections the rate of household growth is slower after 2021. The main factors here are that the number of deaths increases as more of the population is in age groups with higher mortality rates and increased net migration out of Northampton to the rest of the UK, only partially offset by rising headship rates.

5.3.   The 2011-based household formation rate projections are both significantly lower than those underlying the 2008-based projections and rise much less quickly. (It should be noted that, even so, the projected household formation rates do rise, indicating a fall in average household size, but at a much slower rate than had previously been projected.)

5.4.   To understand what is going on it is necessary to delve into greater detail and look at the different age groups and household types.

5.5.   DCLG produce their household projections using 17 household types and 9 age groups. Appendix 1 shows the 2008-based and 2011-based household formation rates for each of the 9 age groups for all household types in the West Northamptonshire area. The charts have been drawn on a similar scale to facilitate comparison. For most age groups the differences between the two sets of projections is small. There are, however, three age groups which stand out: 25-34, 60-64 and 65-74. For these age groups there is both a sizeable difference between the headship rate in 2011 and a noticeable difference in the slope of the line. As it is the change in the number of households that is important in planning for housing, differences in the slope of the headship graphs – implying different rates of change – are particularly significant.

5.6.   Looking first at the 25-34 age group, the chart below reproduces the headship rate graph from Appendix 1:

5.7.   Note that the starting point for the 2011-based projection in 2011 is below the headship rate 20 years earlier in 1991. Note also that, whereas the headship rate of this group had been growing steading since 1991, the 2011-based projection suggests that the headship rate of this group will fall more steeply than it has grown. The key issue is whether this is a realistic indicator of a likely future trend or whether it is the result of the weight that has been given in the projections to the 2011 census results – a census that was taken three years into a severe economic downturn and after a sustained period of poor housing affordability. As such it is plausible that the census reflects a low point in the economic cycle that is not indicative of the likely longer term trend.

5.8.   Further insights can be gained by looking at the different household types in the 25-34 age group. The chart below groups these into 5 broad groups to make it easier to see the pattern.

5.9.   The biggest difference between the two projections is amongst the single person households. This is consistent with young adults either remaining in the parental home for longer or living in shared accommodation with other young people rather that setting up home on their own. On the latter point a group of young people sharing a flat or house with other young people would be classed as an ‘other’ household. The chart shows an increase in the formation rate of such households. That increase is much smaller than the reduction in the number of single person households, but that is to be expected if some of the ‘missing’ single person households are grown-up children living with parents. Moreover, if four 25-34 year olds share a house or flat rather than each setting up home on their own there would be a ‘loss’ of four single person households and a ‘gain’ of only 1 ‘other’ household. This means that the increase in the headship rate for ‘other’ households would be much smaller than the decrease in the single person headship rate even if all of the ‘missing’ single person households had become shared households.

5.10.   There is some further evidence for this from the detailed breakdown of the headship rates for households in the age groups that will include the parents of those age 25-34. For example, the chart below shows the headships rates for the 55-59 age group:

5.11.   A grown up child coming to live with their parents would turn their parents’ household into a couple or lone parent household with another adult as ‘children’ here means dependent children, not adult children. In the chart above a significant increase in such households can be seen.

5.12.   There is also a reduction in the number of couple households in this age group. This is consistent with adult children coming to live with their parents as, when that happens, a couple who might have previous been living on their own as ‘empty nesters’ would move from being a couple living on their own to being a couple with other adults.

5.13.   The reduction in the headship rate for single person households in this age group has a number of possible explanations. It could be as a result of fewer relationship breakdowns or, where a relationship breaks down, the parties having to accept shared accommodation rather than being able to afford to set up as separate households. Another possible explanation for at least some of the reduction in the number of single person households in this age group is children coming to live with lone parents. Such lone parents would be classified as single person households. If one or more of their adult children returns home, that household then becomes an ‘other’ household.

5.14.   It should be recognised that all that can be said here is that the pattern of the changes in headship rates is consistent with more single young adults either living with their parents or in shared accommodation, not that this is conclusive proof. There is, however, other evidence that there has been a significant increase in young adults living with their parents. In May 2010 ONS released “Young adults living with parents in the UK, 2011”[8] . Using data for the Labour Force Survey this suggested that there had been a 21% increase in the number of young adults living with their parents between 2001 and 2011 – an increase of over ½ million people – as shown in the chart below.

5.15.   As can be seen from the chart, it would appear that the growth in the number of young adults living with parents started well before the credit crunch and recession.

5.16.   Whilst it is possible that some of these behavioural changes will have been free choices, it seems more probable that most are changes caused by the economic situation, the cost of housing and the difficulty in obtaining a mortgage without a sizeable deposit. It would be prudent to plan on the basis that there is a likelihood that household formation rates in this age group will move at least some of the way back towards the previous trend if economic conditions improve.

5.17.   What is happening in the 60-74 age group is less clear. The chart below compares the 2008 and 2011-based household formation rate projections for the 60-64 age group. (The 65-74 age group is similar.)

5.18.   The difference in household formation rates in 2011 is not as great as for the 25-34 age group although the difference in the slope of the lines is as dramatic. The chart below breaks the age group down into the same five broad household types used above:

5.19.   The reduction in couple households and increase in couples and single parents with other adults could be the result of adult children coming to live with parents, although the reduction in couples is larger than the increase in ‘couples and single parents with other adults’.

5.20.   It is possible that some of the reduction in single person households could be the result of grown-up children coming to live with single parents, but the increase in ‘other’ households is too small for this to account for more than a fraction of what is happening here.

5.21.   Looking at the headship rates for this age group in 2001 may offer some clues. The chart below adds that data in the grey bars.

5.22.   It is noticeable that the single person headship rate in 2001 was very similar to that suggested by the 2011 based projection (whilst the changes in ‘couples’ and ‘couples/single parents with other adults’ are consistent with grown up children moving in with parents). It may simply be that the 2008-based projections were too bullish about the growth in headship rates for single person households in this age group.

5.23.   It should also be noted that the 60-74 age group is much less likely to be affected by the economic downturn as they will have been well on in their ‘housing careers’ before the downturn occurred. Indeed many homeowners will have had substantial equity in their properties.

5.24.   Given the significant uncertainties in the 60-74 age group; the possibility that at least some of the differences between the two projections are due to over optimism in the 2008-based projections; and the likely lower impact of economic conditions, it is far less clear that there will be a return to the 2008-based trend in this age group. On the other hand, a continuation of the fall in household formation rates projected for 2011 to 2021 is also very far from certain.

6.5. As can be seen, for the West Northamptonshire area the ‘part return to trend’ and ‘2008 tracking’ options are virtually indistinguishable. (Indeed, markers have had to be added to the data points for the ‘part return to trend’ option to make it visible.) This need not necessarily be the case: it is merely a result of how the relative sizes of the populations in the different age groups have combined with the differences in the headship rates in each group to produce a very similar overall effect.

6.6. Given the closeness in the overall headship rates it is unsurprising that the two scenarios produce very similar household growth projections, as shown in the chart and table below.

6.7.   The ‘Partial return to trend’ and ‘2008 tracking’ scenarios both imply an average annual household growth that is some 10% below the 2008-based projection compared with the 2011 extended scenario that is some 20% below the 2008-based projection.

6.8.   There is a sense in which both ‘no return to trend’ and ‘full return to trend’ would be extreme scenarios. If one or the other were adopted as a planning assumption there would be a significantly higher risk that there would need to be a sizeable correction part way through the plan period as a result of either too many or too few homes having been provided for.

6.9.   By following one or other of the (virtually indistinguishable) partial return scenarios the risk of needing to make a significant adjustment would be reduced, although regular monitoring of what actually happens would still be important.

6.10.   When converting the household growth projections into an estimate of objectively assessed needs allowances need to be made for vacant and second homes and existing unmet housing need.

The 2011 census-based population estimate for 2011 was 2.4% below the 2008-based population estimate. However, the projected growth in the population over the period of the official projections (2011 to 2021) is 9% faster in the 2011-based projection than in the 2008- based projection.

Drivers of population change

The following chart summarises how the projections for births, deaths and migration flows are estimated to drive population change in the three population projections:

Household formation rates

Chart below shows how the 2008-based and 2011-based household formation rates for Northampton compare.

Note that the 2011-based household formation rate for 2011 is both lower than the 2008-based projection and increases much more slowly.

Note also that the population growth up to 2007 is somewhat uneven. A possible explanation for at least some of the ‘dip’ between 2001 and 2007 is that 1500 homes were built in South Northamptonshire as an extension to Northampton. At an average household size of 2.4 people that would equate to 3600 people living in South Northamptonshire who were, in a real sense, part of the growth of Northampton, but who will have been counted (quite correctly) as South Northamptonshire residents. ‘Adjusting’ the Northampton population growth figures to include the extension in South Northamptonshire would go some way to filling in the ‘dip’. (As an indication of scale it might be noted that between 2001 and 2007 the population of Northampton grew by 11,000.)

As the chart below shows, at a similar time there were ‘bulges’ in the population growth of both South Northamptonshire and Daventry:

When the three authorities are combined, as shown in the chart below, a much smoother population growth curve emerges:

This illustrates two general points:

• Population (and household) growth can be affected by the number and location of homes built, with growth in part reflecting where homes are built.
• Local authority boundaries are not particularly relevant when considering how housing markets work. A major advantage of producing a joint core strategy such as that for West Northamptonshire is that the needs can be addressed in appropriate locations across the sub-region.

As far as the impact on the population growth projection for Northampton is concerned, the slower growth between 1997 and 2005 does not appear to have had a large impact on either the 2008 or the 2011-based projections. It is of course true that how population growth is actually shared between the three authorities will depend in part on where the homes that are needed are built. It will also depend on the use of vacancies and densities within the existing stock.

Household growth

The chart and table below compare the 2008 and 2011-based household projections for Northampton.

The 2011-based projection suggests household growth over the period 2011-21 that is 14% slower than the 2008-based projection.

The 2011 census-based population estimate for 2011 was 2.3% below the 2008-based population estimate. The projected growth in the population over the period of the official projections (2011 to 2021) is 12% slower in the 2011-based projection than in the 2008-based projection.

Drivers of population change

The following chart summarises how the projections for births, deaths and migration flows are estimated to drive population change in the three population projections:

Household formation rates

Chart below shows how the 2008-based and 2011-based household formation rates for Daventry compare.

Note that the 2011-based household formation rate for 2011 is both lower than the 2008-based projection and increases more slowly. However, the differences are smaller than for Northampton.

Household growth

The chart and table below compare the 2008 and 2011-based household projections for Daventry.

The 2011-based projection suggests household growth over the period 2011-21 that is 27% slower than the 2008-based projection.

The 2011 census-based population estimate for 2011 was 6.5% below the 2008-based population estimate. The projected growth in the population over the period of the official projections (2011 to 2021) is 19% slower in the 2011-based projection than in the 2008-based projection.

Drivers of population change 

The following chart summarises how the projections for births, deaths and migration flows are estimated to drive population change in the three population projections:

Household formation rates

Chart below shows how the 2008-based and 2011-based household formation rates for South Northamptonshire compare.

Note that the 2011-based household formation rate for 2011 is both lower than the 2008-based projection and increases more slowly. Again, the differences are smaller than for Northampton.

Household growth

The chart and table below compare the 2008 and 2011-based household projections for Daventry.

The 2011-based projection suggests household growth over the period 2011-21 that is 29% slower than the 2008-based projection.