As part of the Changes to the current planning system consultation, MHCLG has produced a new standard method for assessing housing need (SM2). The method produces an overall need for over 337,000 homes per annum across England.

Savills previous work on the Standard Method was acknowledged in the consultation. We discussed our approach with MHCLG and welcome the changes now proposed.

Like the previous method, SM2 views affordability as a driver of housing need, with worse affordability indicating a lack of supply against demand. Affordability is used to adjust a baseline housing need generated from household projections or 0.5% of existing housing stock, whichever is higher.

A step in the right direction…

The proposed SM2 is a good step forward the formulaic calculation of ‘policy off’ in housing need. It introduces housing stock as part of the formula, reducing the influence of household projections. This should lessen the impact of the feedback loop inherent in household projections, that places with low levels of past housebuilding are projected to need less additional housing in future.

Although a good step forward, we think the proposed SM2 is still overly reliant on household projections to generate housing need figures. So it does not solve problems such as the low projected population growth in cities, highlighted in our previous paper.

…but could go further

The housing need result is still calculated to be relatively low in some of the least affordable places in the country.

The maps below show that this is most obvious in the local authorities surrounding London. Some of these have proposed housing need figures well below the national average as a percentage of stock. This is because 0.5% of dwelling stock is too low to address suppressed household formation in the least affordable locations.

It continues to allow local authorities who have historically failed to meet need, so have low household projections, to continue planning low numbers of new homes.

Some of the highest housing need is in parts of the Midlands, where there have been high levels of recent housing delivery. So SM2 does not fully address the household projection feedback loop referred to above.

Stability lacking

Use of the 0.5% of dwelling stock is also intended to bring stability to the housing need figures. But only 25% of the total baseline for England is based on dwelling stock rather than household projections.

The change from 2016- to 2018-based projections would have resulted in a 20,000 home fall in housing under SM2, the same as under the current system. The fall between 2014- and 2016-based projections would actually have been larger under SM2.

Still a ‘starting point’… for now

SM2 remains a housing requirements in Local Plans. Any additional housing needed as part of regeneration and renewal or to accommodate additional demand generated by employment growth needs to be considered in addition to these numbers. Where these issues dominate, it should not be surprising that Local Plans include housing requirements that go beyond a housing need figure based only on demographics and affordability.

The Planning White Paper promises a Standard Method 3 (SM3), which will be binding on local authorities and much more complex. In addition to affordability, it is proposed to have regard to the size of existing settlements, the extent of land constraints and the presence of brownfield land.

Where the baseline applies

The affordability adjustment is applied to a baseline number generated from household projections or 0.5% of existing stock, whichever is higher. Those using the dwelling stock baseline are shown in red on the chart.

Local authorities using 0.5% of dwellings as a baseline figure produce housing need figures that are not much higher than the national average even in the least affordable places.

The highest housing need as a proportion of stock using dwellings as a baseline is in Epsom and Ewell at 1.9% of dwelling stock. In 2019, house prices in this local authority were over 16 times earnings. The baseline figure in Epsom and Ewell required an uplift of 3.17, the ninth highest uplift applied under SM2.

Baseline rationale

The rationale for using a baseline derived from dwelling stock is to ensure it is high enough in places where household formation has been suppressed by poor housing affordability. We would therefore expect the 0.5% of stock baseline to have a greater influence in less affordable areas.

Actually the opposite is true: the baseline is more often in use in the most affordable locations. It is not clear what the rationale would be for increasing the baseline beyond demographic projections in these areas.

A better approach?

Savills proposed method ensures that dwelling stock-based housing need is most used in less affordable places, as shown below.

Taking Epsom and Ewell as an example, our proposed method would require an increase of 2.9% to dwelling stock per annum.

Under our proposal, the most affordable locations are not affected by a dwelling stock-based floor as there is little evidence in these locations of suppressed household formation. The housing issues in these areas are not affordability based and the overall shortage of homes is less acute. However, there may be complex challenges such as poor quality housing, which a simple formula cannot address but would presumably be tackled by SM3.

Adjusting the affordability adjustment

In SM2, the affordability uplift is made up of two parts: one simply based on the level of affordability, and one based on the change over 10 years.

These are shown below, with a point for each local authority showing its uplift from each component. Twice as much of the adjustment derives from the change in affordability over 10 years compared to the current level of affordability.

The rationale for using the level of affordability in the housing need formula is clear. It is a proxy for the mismatch between existing housing stock and demand for homes.

The rationale for using the change is less clear. If affordability has improved but homes remain relatively unaffordable, there is no rationale for lowering housing need.

Our analysis shows that the two components are strongly correlated anyway, so a simpler formula with a higher adjustment based on current affordability could achieve a similar result.

Affordability instability

The new affordability component in SM2 is the use of the change in affordability ratio over a 10-year period. This is likely to introduce additional variation to the housing need results.

We have tested this by running SM2 on historical household projections and affordability numbers. The results are shown in the table, below.

The variation over time is substantial and is driven by the change in affordability over time, which in turn is susceptible to the cyclical changes in house prices.

The 2019 uplift based on change in affordability is particularly large, as the beginning of the 10-year window (2009) was when house prices bottomed out following the global financial crisis.

Over the last three years, the affordability ratio has been fairly stable in many places. If this continues then the 10-year change to 2020 will be similar to the change between 2010 and 2019. Using this nine-year change would cut housing need by over 20,000 homes, down to 315,700.

The Liverpool inconsistency

The revised affordability adjustment also produces some unexpected results at local level.

In Liverpool it results in a housing need figure lower than the projected growth in households, as the current house price-to-earnings ratio is 4.22 and the ratio has fallen by 0.33 over the past 10 years.

This suggests that Liverpool actually has an oversupply of homes despite having an affordability ratio greater than the proposed threshold of 4. This seems inconsistent.

The importance of 4

Applying the affordability uplift to anywhere with a ratio greater than 4 is a key decision, justified in the consultation simply with reference to the average first-time buyer loan-to-income ratio in 2015 (3.61). This is cited as a finding of the Council of Mortgage Lenders (now UK Finance).

The significance of 2015 is a mystery. The significance of that specific ratio is not much clearer.

The ratio used in the SM2 formula is the median house price-to-median earnings ratio published by the ONS. The house price in that ratio is the median price paid of all residential sales recorded by HM Land Registry. The earnings is based on the ONS Annual Survey of Hours and Earnings and is the median amount earnt by a full-time worker employed in an area.

The mortgage loan-to-income ratio cited is different. Most importantly it is only the loan that is being compared to income, not the house price. UK Finance data allows the average first-time buyer house price-to-income ratio to be calculated. It was 4.5 in June 2020.

The income number is also different to the one in the ratio used by SM2. It is the income (earned income and possibly other income) of the borrower. So it may include joint incomes. And it is only the incomes of those people affluent enough to be buying houses.

Although the amount people can borrow is limited, the greater barrier to home ownership is the deposit requirement. The average first-time buyer needed to raise 22% of the purchase price in 2019, equivalent to 101% of their annual income, according to UK Finance data.

The Government's First Homes scheme aims to address this, but will only help a relatively small number of people.

Widespread availability of high loan-to-value mortgages seems unlikely to return in the short term. Increasing availability of such products over the last few years has recently been largely reversed by the economic uncertainty stemming from the Covid-19 pandemic.

Many households, particularly those on lower incomes, will struggle to save a sufficient deposit to buy a home and are likely to be long-term renters. The existence of both private and social rented housing is likely to continue, and in general, is likely to accommodate less affluent households.

This all suggests that the threshold should be higher than 4. Our proposed approach introduced a floor to housing need, based on dwelling stock and affordability, only in places where the median house price-to-earnings ratio is greater than 5.