Joint Core Strategy - Proposed Main Modifications

Response from Peter Brett Associates to JPU queries



Question 1

The first issue we have is with the results specified in para 5.3. These numbers, especially for South Northamptonshire, but also to some degree Northampton, look too low. It would not be possible for us to accept these even as minimum targets without a much better and clearer justification.

Answer 1

For the JCS area as a whole:

  • The numbers are our best estimate of the ‘objectively assessed need’ that the NPPF requires local planning authorities to meet. We are not suggesting you accept them as a target, not even a minimum target. Rather, we are saying that an NPPF-compliant target cannot be lower than this number. But it may be higher, even much higher. We can expand this point in the report. We could also discuss the many reasons why you might want a higher target.

For the individual local authority areas:

  •  Again, we are not suggesting you use the projections as targets. We do think the projections can help inform your targets. But at this level they may not be good indicators of demand or need, because much of that demand or need may be footloose across local authority boundaries. In particular, many people who want to live in or around Northampton town would be just as happy with a house which is part of the Northampton urban area, but technically in one of the other districts. The projections should be interpreted in the light of local knowledge – understanding where people come from, where they want to be and why.
  • Thus interpreted, the projections for each district give us a first indication of where the demand may be (where people would like to live). But the projection cannot tell us in which district this demand should be accommodated. These decisions depend on supply capacity and policy, not just demand. For example, the authorities may choose to accommodate some or all of the demand projected for Northampton on the edge of the town, one of the other districts. The projections cannot tell us whether this is a good idea. It depends where the deliverable and sustainable supply capacity is.
  • There are other reasons why your targets may be very different from the projections.

    These may relate to future population profiles. Thus, in South Northants, if your target was based on our Trends 2001-11 scenario, the population would grow slowly and age fast: the projections suggest that nearly all the population growth would be at ages 70+ (Appendix, Table 8 and Figure 10). This could have bad implications for the sustainability and vitality of certain settlements.

Question 2a

The 50 per annum for South Northamptonshire in not at all consistent with actual delivery over the past 5 years, let alone a longer period.

Answer 2a

The demographic projections do not take account of housing delivery. All they do is roll forward demographic and household formation rates. Thus, let us assume hypothetically that in the past (base period) 10% of 79-year-old males died each year. If in the future the number of 79-year-old males doubles, then other things being equal the number of deaths will also double. The above is a caricature, or extreme version, of what happens in our projections for South
Northants (Appendix, Figure 10). In Trends 2001-11, and even more so in Trends 2006-11 (which excludes the shot in the arm from Grange Park), there is little inward migration and the existing population ages dramatically. Therefore there are many deaths and few births, little or no population growth, and correspondingly little household growth.

In short, demographic projections show the impact on population and household numbers of changing age profiles. In the case of South Northants, these impacts are large.

Question 2b

In addition [the projection for South Northants) does not square well with the 900 plus population increase shown in the recent 2012 Mid Year Estimates.

Answer 2b

The S Northants increase in the 2012 MYEs is surprising, far above what one would expect from recent years. We do not know if special factors account for it or if it is an error. Either way, no one would project forward 20 years on the basis of one year only. The ONS’s next set of projections will use the average of the five years to 2012, which will make around 200 p.a., similar to the 180 p.a. in our Trends scenario.

Question 3a

From our reading of the document and other research we have done, it would appear that this low figure may be related to the -3,500 population adjustment in the “Other” category, which appears to reflect the over-estimation in the 2008 based figures, and then projected forward. We are not sure that this is appropriate.

Answer 3a

The adjustment is unexplained (or unattributed) loss, i.e. population change which ONS cannot account for. That is, the Census population counts show that this change has happened, but records of births, deaths and migration do not say how it has happened. Population counts and births and deaths records are highly accurate, whereas migration data are not. Therefore the likely cause of the problem is bad migration data.

These unexplained changes are included in our projections. In our opinion this is appropriate, because the changes did happen. However, we expect the new ONS / CLG projections will exclude them, because their models can’t deal with them (it’s much harder for the whole country than just W Northants).

If unexplained changes were excluded from our projections, the projected household change across the JCS area would go up slightly, perhaps from 1,500 to 1,800 p.a. But to make calculations that are wrong on purpose would be unhelpful.

Question 3b

A similar issue applies to the Northampton data also, although in this case the 2012 MYE is significantly below that forecast in the 2011 based projections.

Answer 3b

Again, one would not base long-term projections on a single year.

Question 4

On a broader front, there has been views expressed in the technical press that
      i. The perceived under-enumeration in the 2001 census could result in inter censal change rates being distorted. If appropriate a comment on this and any  implications would be useful.

Answer 4

We do not know of any evidence that the 2001 Census was more undercounted than 2011. We do know that ONS tries hard to correct any known undercounts and other errors in the Census. See for example q=ONS+census+undercount&oq=ONS+census+undercount& .

If, hypothetically, 2001 was more undercounted than 2011 and not corrected, then the official statistics would overestimate the growth in 2001-11 (start point too low and end point correct). New, corrected projections would show lower growth than our existing projections.

Question 5

ii. [Also] that any rates and projections based on the 2006-11 period could be argued to be “locking in” the recessionary period, and therefore generating low forecasts, particularly in the medium and longer term. Again, a clear discussion of this and the implications/consequences would be useful.

Answer 5

I think this comment came from me (in Planning magazine).

CLG’s latest housing formation rates, which we have used, are based on the four last Censuses, from 1971 onwards (plus data from the Labour Force Survey). They lock in a long-term trend that includes one very big housing recession in 40 years. This latest recession was exceptional. In relation to housing, nothing like it had happened in the previous 40 years.

One could take a more optimistic view, that the current recession is an exceptional one-off event, demand will recover from it quickly (including due to recent government action on mortgages, housebuilding etc) and nothing like it will recur during the plan period. As I said to Planning magazine, ideally the CLG would provide projections on that basis – perhaps as a variant or sensitivity test. Since CLG have not provided that alternative view, I think it would not be
helpful for West Northants JPU to do so. It would take you too far from the official projections, see Question 6 below. More important, it would take you away from what the JCS Inspector asked for, see Answer 6 below.

Note added on 3 December 2013: Since this was written, the South Worcestershire Local Plan Inspector, in his Interim Conclusions[1] , has endorsed a projection in which ‘current trends in household formation will persist until 2021, after which there will be a return to the household growth rates experienced in the years before the financial downturn’. This approach is similar to the ‘more optimistic view’ described in the last paragraph. This suggests it might be reasonable for West Northamptonshire to take an alternative, more optimistic view of future household formation.

Question 6

You have argued in 2.14 that less weight will be given to projections run by local authorities….. consultants etc, yet you have done precisely that, based on rates which you have created yourselves.

It seems to me therefore, that substantially more justification must be provided to support the use of your own projections as opposed to those of the government, especially where the outcomes are count-intuitive based on historic data.

Answer 6

Weight given to our projections

2.14 says that the official projections should be our starting point and alternative projections carry less weight – not ‘nil weight’. On second thoughts, I would rephrase it to say that the official projections are an essential starting point and authorities should not depart from them without good reason.

In this case, there are such good reasons. Projections based on arbitrary assumptions are not helpful. But you may, and indeed you should, correct projections that are based on wrong facts. Projections are nothing more than past facts rolled forward. If these past facts are wrong, the projection is not credible.

In this case, this is exactly what the JCS Inspector asked. His request was that you update the official projections to take account of the 2011 Census. As it happens, that Census showed that the ‘facts’ on which the official projections were based were badly wrong.

So we provided alternative projections that correct for these errors, and in every other respect follow the ONS / CLG methods as closely as possible. The rates we use are not arbitrary creations. They are derived by these methods, which roll forward past trends and explained in the report.

In summary, we consider that our projections carry more weight than the official projections because they are based on corrected historical facts, and the corrections are very large. If you disagree, you should simply note our advice and revert to those official projections.

Our approach has been as set out in our proposal to you. We cannot think of any other way to fulfil the Inspector’s request or of any further justification.


In the light of the historical data, as revised in the light of the Census, our projections look intuitively plausible. Thus in our preferred scenario, Trends 2001-11, future household growth for the JCS area is 1,540 p.a. - close to 2001-11, when it was 1,400 p.a.. (The official projections are based on CLG’s previous estimate of this past growth as 2,390 p.a. – almost double the actual figure.)

For smaller areas and shorter periods, the future is more different from the past, because change is more volatile. As regards small areas, see earlier comments, especially about South Northants. As regards short periods, as noted in the report we do not set much store by Trends 2006-11. If you prefer, we can give it even less weight in the report or even leave it out altogether.

Question 7

It would also be useful for us to have some of the interim output and components of change details, as this will enable us to look further into how your figures are compiled.

Answer 7

Some details are in the Appendix. Full details are in the working Excel file attached.

1. [back]