As many investors continue to debate over whether the property market is going through a process bottoming out, one of the main challenges that continues to face the industry is a seeming lack of concrete housing data to undertake truly valid analyses - largely due to the fact that there are such low levels of sales activity. We are therefore very pleased to present a very interesting an insightful interview with David Thorpe of Acadametrics – a consultancy that provides detailed housing-related information to a range of important institutions and is regularly featured in broadsheet media sources. We talk in detail about how the organisation undertakes housing research; the UK house price index system and its related criticisms; stress-testing the UK housing industry; the current fragility of the property market and some thoughts for the future.
1) Who are Acadametrics? We are a privately owned housing analytics company which has worked in the sector from 1988.
2) Who are the clients that you mainly work with? Banks and building societies but surveying and estate agency firms, hedge funds, property funds – indeed, with any organisation which is active in housing.
3) Can you talk through the major reports and statistical analyses that Acadametrics produce? Our flagship report is our monthly News Release and associated information, published on our website, reporting the house price trends shown by the LSL Property Services/Acadametrics House Price Index (LSL Acad HPI). In addition, we publish on our website e.g. our House Price Calculator (HPC) enabling a known past value of a house to be updated in accordance with the data underlying our index: this is similar to the calculators provided on the lender websites but the updates are calculated at a more granular level, subject to the availability of adequate data.
4) The majority of our readers are property investors and developers – and, as you mention on your website, the main difficulty that people are having in the current climate is researching the true market value of property. What are your thoughts / advice? We will not need to remind your readers that the property market is a market like any other and a house is worth whatever a willing buyer is ready to pay. An index, such as the LSL Acad HPI, provides a measure of monthly and annual house price inflation and an average price for a particular geography at a particular time. Indices are backward looking, as are house price calculators. Even our HPC provides only a broad brush estimate, since it uses inflation by property only at county/London borough level and may have to use inflation at regional level if not enough transactions take place of a particular property type within a particular county for reliability. A full explanation is linked to our statistical HPC. An index and an HPC are amongst the tools which can be used by property investors and developers. We also provide the Acadametrics Prices and Transactions (APAT) data used by our HPC as a monthly series, showing the average price and transaction numbers for detached, semi-detached and terraced houses and flats for each county and London borough. APAT data are used to reveal trends in market values. We are adding to APAT optional monthly data at postcode district level; in addition we offer forecasts of prices for the four property types at postcode sector level, using a “quant” procedure developed by analysts TFL whose housing sector work we have taken over. Please see our comment on house price forecasts below.
5) Can you run through the essential stages you adopt when compiling your analyses? On the first working day of each month, the Land Registry send us the average price and transaction numbers by county and London borough for all the transactions registered on their database for the immediate past and all prior months. We use the data for the immediate past month to prepare the current house price index % Monthly inflation, % Annual inflation and average price results for England & Wales, each region and each county/London borough, together with our News Release and many associated reports and data tables. LSL Property Services PLC add a practitioners commentary to the News Release which is then sent by the Wriglesworth Agency to the media and are shown on our website. Using the prior month’s data, we update the inflation and average price results provided in our earlier releases until additional transactions have no further influence on our results. At this point, we describe the particular results as LSL Acad HPI “ultimate”. We will freely send LSL Acad HPI every month to readers who care to send us an email address through the link here.
6) What do you think are the most trustworthy indices in operation in the UK? Every Index is trustworthy in that we may assume that each provider makes the best possible estimates based upon the available data. We would certainly expect developers and investors to compare and contrast the indices. Our Index Monitor, published on our website, tracks how closely or otherwise each index compares with the inflation results shown when most transactions have been reported and prices are approaching “ultimate”. Readers might well want to use Index Monitor to establish how much confidence they wish to place in a particular index.
7) You have partnered with the ‘Zoopla’ organisation – why did you choose this particular index of house prices to work for? AS well as the property website with which readers will be familiar, Zoopla provides an Automatic Valuation Model (AVM). When our statistical HPC shows that APAT revalues a property only with say two standard deviations, when revaluing a lender’s residential property portfolio, we would recommend that an AVM be used for such a property. We chose the Zoopla as being an excellent AVM, in our opinion.
8)The National Statistics Authority is planning to investigate the house prices indices and I wondered if you could provide a response to this? We would ask to be excused from providing a public response before we have answered the ONS questionnaire. Suffice to say that we have already queried, at a Royal Statistical Society meeting, the existence of two official house price indices and three official house prices. Readers are likely to be aware of – but many people will be misled by – the concept of a “standardised” average, as opposed to a simple average, house price. A standardised average, as provided by the Nationwide and Halifax indices, takes the average calculated at a past date and updates it monthly according to the movement of the index. In the case of the lenders, the past price was calculated hedonically from the values assessed, at the time, of e.g. a bedroom, a bathroom and a garage. In the case of the Land Registry (which lacks details on numbers of bedrooms and bathrooms etc), the standardised average is that calculated at 2000, backdated to 1995 or updated from 2000 by the value of the index. CLG (like our own index) calculates a mix adjusted average price using current prices. Land Registry also calculates a quarterly simple average price which they provide for the BBC website; hence, the existence of three official average prices at national level, for February 2010, as follows:
- Land Registry simple average £224,064 (taken as a February price from the BBC Q1/10 average)
- Land Registry standardised average £164,455 (taken from the February Land Registry index)
- CLG mix adjusted average £204,359 (taken from the February CLG index)
Readers, but not all of the public, will be aware that the: lenders use prices taken from each surveyor valuation at the time of a mortgage offer; CLG uses prices agreed at the time of each mortgage completion; Land Registry uses the reported transacted prices (but, for the index, only the circa 35% for which a price from a prior transaction is available). Also lender and CLG prices are for the UK; Land Registry prices are for England & Wales. Be these details as they may, our concern is that the public may not be able to see any reliable house price wood amongst all the house price trees. For the official statistics to report that the average house price is £224,064 – alternatively £164,455 – is confusing. Average house prices reported by CLG are consistently quite close to those which we calculate.
9) You have many years of stress-testing various aspects of the UK housing industry - can you explain the various aspects of how you work in this regard? We take the lender’s portfolio, or the part of it upon which the lender wishes us to work, and update the value of the properties using our APAT data. At portfolio level, our tests have shown that APAT provides a very closely accurate value. Using our look- up tables providing the “hazard” or probability of redemption or possession for each property, based upon the performance of a comparable group of properties during e.g. the 1989-1991”worst case” housing crisis, we the estimate a Probability of Possession and a Loss in the Event of Possession for each loan. Our database enables us to prepare the above under different macroeconomic scenarios including “current” and the FSA ”worst case”. With our New York colleagues MIAC Analytics, we now offer users the ability to undertake this work in-house using the Acadametrics and MIAC data and software on the MIAC DataRaptor OLAP database management engine and MIAC WinOAS cash flow management tool, downloaded from our secure UK server.
10) Can you describe what the “Stress and Scenario Testing for UK Residential Mortgage-Backed Securities; The Requirement for Loan-By-Loan Testing” document is and why it should be read by those in the housing industry? Lack of funding is a key issue in today’s housing market. A revival of securitisation for residential loans would provide a part if not a full solution. Investor demand and, hence, investor confidence is, in turn, key to a revival. Our Dr Stephen Satchell (Economics Fellow Trinity College Cambridge) wrote this Discussion Paper for MIAC ACADAMETRICS LTD to describe the movement in the USA and EC towards loan level stress testing as a means to secure investor confidence. We recommend Dr Satchell’s paper as valuable reading and a guide to the future of the market for developers and investors.
11) On a general level, how fragile is the UK housing market at the present time? The UK market is a regional and even a local market as readers will know. The RICS survey is required reading for professional opinion as to local trends. At national level, LSL Acad HPI shows average house prices virtually unchanged from £221,074 in April to £220,685 in July. Whilst the July price is our forecast based upon the c.35% of transactions reported to Land Registry by month end and will be subject to change when many more July data are available at end August, our prices are true averages (smoothed over three month periods), as opposed to the “standardised” prices reported by the lenders and the Land Registry index. Be that as it may, the trend is flat and, to the extent that it could go in either direction, we would certainly say that, even at national level, the market is fragile. Our APAT will indicate local trends.
12) Can you provide us with your prediction on the short (1 year), medium (5 years) and long term (10 years) movement of house prices – and how you have come to this prediction? Readers will understand that only Paul the octopus, fresh from success at the World Cup, would provide a confident house price prediction. Forecasting models, whether based upon macroeconomics, or (such as ours) based upon quantitative analysis, are thrown off course by events such as the financial crisis from which Western economies have not yet emerged. The huge house price falls which this caused were mostly unforeseen and many economists were equally surprised by the upturn which occurred in Spring 2009 but which is now flattening out. We provide a forecast only as a trend line to show how we would expect prices to progress, given no sudden changes to general conditions, such as in the economy or the legal framework around house purchase. Our “quant” forecasts use a very large number of economic data series and neural statistics to show how prices should move given anticipated movement in the data employed. Such a trend line should not be used to provide a definitive forecast of prices in a particular postcode sector, or other particular geography, at a particular date, but to provide a long term benchmark trend, against which to measure the factual changes which are occurring. With such a benchmark (and others, such as the much used price to earnings ratio), it is possible to make better judgements as to e.g. by how much prices are above trend (and if they are in a bubble), or below trend, and what changes might thus be expected to take place. Since our forecasts have the above caveat and particular purpose, readers will understand, we are sure, that we do not place them in the public domain lest they be misunderstood.
