By employing a variety of econometric methods many previous studies have concluded that there is long run convergence present between house prices in the South East of England and other regions within the UK. Using quarterly data for the period 1974:1 - 2009:4 this study uses the ripple effect hypothesis to analyse the strength of convergence in each region of the UK through the application of unit root testing procedures, argued to be the advantageous approach above other methods such as univariate. It considers how the average South East house price affects house prices elsewhere, and analyses possible explanations for this result.
Key words: Ripple Effect, convergence, cointegration, weak segmentation, regional house prices
During the past 30 years there has been extensive studies concerning the price of housing in different regions of the UK and whether or not these prices are related to each other. Unsurprisingly there have been great variations in ideas and opinions of where, why and how these price differentials may exist. It has become clear from (most) studies that there is a convergence of prices throughout regions of the UK. Note that convergence in this context refers to a continuous stable relationship or equilibrium between prices of one region to another, as opposed to that of an equalisation of prices over a period to a specified point. From previous studies it has been indicated that housing prices in the South East of England tend to adjust more quickly and more violently than any other region. Price fluctuations in this region then in turn affect the prices of houses in other regions of the UK over different periods of time and by different ratios. This movement in prices has become commonly known as the "Ripple Effect". This study aims to add to previous academic writings by interpreting between which regions the strongest convergence rates can be seen, and possible explanations for this.
Foremost this study holds a very strong practical importance as house prices influence the majority, if not the entirety, of the population. Housing prices not only affect where people ultimately live, but this in turn can directly affect businesses and the economy due to a lack of labour mobility causing shortages of skilled labour in those regions with relatively high house prices, such as the South East.
The academic interest in this topic over the previous decades has been consistent and there is still active interest in the differential of house prices to date. Central to the most recent work of housing economics has been Steven Cook who has used the application of many econometric tests to analyse the convergence of the UK housing market, publishing many papers over recent years. The most predominant earlier workings came from the authors Nellis & Longbottom (1981), whose work remains significant today, as mentioned in many articles.
Throughout the literature there have been many controversies surrounding the issue, although less over recent years. The most predominant is that of whether the infamous "Ripple effect" ultimately exists, and if so the strength in which it holds is questionable. Those that claim to have produced evidence of this phenomenon existing are Meen (1999) and Cook & Thomas (2003), however there are others that somewhat disagree with the theory existing such as Ashworth & Parker (1997). Due to inconsistent writings and no solid evidence to show that the ripple effect is always in existence, there is no general consensus as to why the differential of house prices between UK regions exists.
An obvious problem with collecting and analysing data on house prices in different regions is that it is increasingly difficult to compare like with like housing, as highlighted by Forrest (1991). Neighbourhood characteristics, safety, accessibility and many other features are hard to identically compare, causing possible distortions in results which may be hard to detect.
Using data from the last 30 years the aim of this research is to detect in which regions there is the strongest evidence of convergence and at which points in time this has affected the price of housing most significantly. Analysis of these results will then be continued to conclude whether there is any common characteristics which influence the strength of the ripple effect, if at all. The UK data will be divided into 14 regions comprising of: North, Yorkshire and Humberside, East Midlands, West Midlands, South West, Northern Ireland, Wales, Outer South East, North West, East Anglia, Outer Metropolitan, London, Scotland and the UK (see Appendix B).
Intuition may lead one to believe that the strongest convergence rates may be seen in those regions geographically closest to the South East, however Cook (2003) concludes that according to recent findings, North and Scotland have the most "significant evidence of convergence", suggesting otherwise.
Literature Review (3000 words)
Studies on house prices have been carried out for many decades on both a national and regional level. Econometric and empirical analyses have long been carried out with reference to the pricing of housing in the UK with main papers including that of Nellis and Longbottom (1981) in which it was suggested that the rise of house purchase finance encouraged the rapid acceleration of house prices between 1970 and 1980 (one of many papers written by these authors) and Hendry (1984) also carried out econometric work on UK house prices. Both of which focused on the determinants of prevailing house prices of the country as a whole and had great impact on the study, mentioned in many of the following papers by other academics. More recently there has been a much higher interest in house prices at a regional level and many writings on the possible convergence of these prices. More recently Cook (2005) has made a big contribution to the area of study.
There has been concern about whether there is a growing gap in the price differences of north and south. Drake L. (1995) highlighted the issue of the price gap between different regions of the country creating "two nations", as the first author to formally research the topic of convergence (or lack of) between UK regional house prices. There is differing opinion on whether the price gap changes are cyclical or structural. Perhaps gap tends to widen in price booms and shrink in downturns. Closely related to this concept is the well known "ripple effect" hypothesis. In the housing market it is thought that changes in house prices in the South East of England lead the prices of housing in other regions after certain time lags. It is also thought that the prices in the South East fluctuate before any other regions prices, and more dramatically, signalling the direction of the market.
Drakes study shows empirical evidence to support the theory that South East leads the way with house pricing however more recent studies show that this trend has been lost between most regions since the early 1990s (see appendix A). Both studies by Drake (1995) and Giussani & Hadjimatheou (1990) are advocates of the theory that the price gap between the SE and elsewhere tends to widen in price booms and narrows in downturns, which evidence still supports. MacDonald & Taylor (1993) claim to have evidence to support the ripple theory, however an earlier study by Rosenthal (1986) concluded that "London and the South East fail to demonstrate fully convincing and unambiguous evidence of being the driving force behind regional house price diffusion". Both Drake (1995) and Forrest (1991) point out issue of lacking labour mobility due to house prices with issues in SE. Drake uses time varying parameter estimation to test for the convergence between regional house prices, concluding "no evidence is found in this paper for any stronger ripple effect....between house prices in the South East region and elsewhere. Forrest (1991) interested in the significance of house prices in different regions due to the problems with calculating found that 15% of the gap in house prices between South East and North West was due to a poor attempt at comparing like with like houses (distortion) using a regression analysis, highlighted the problems with calculating this issue. Although there is no question as to whether the average house price is considerably higher in the South East of the country Forrest creates a slight contradiction in his writing by first stating that houses are hardest to sell in the South East due to the high prices but then continues to say that this particular region is used as yardstick as the most houses were sold in that area during the period studied. It would be reasonable to use the South East as a yardstick for measurement however as many authors use this region to show the ripple effect.
Cook & Thomas (2003) offered an "alternative approach to examining the ripple effect of UK house prices" by employing a detailed two stage method comprising of both non-parametric and business cycle dating techniques. Carried out after reviewing the differing outcomes of previous works in the field by authors such as Drake, Meen, Ashworth & Parker and Macdonald & Taylor, Cook & Thomas conclude that "strong evidence in favour of the ripple effect was detected". The literature available to date delivers only mixed evidence as to whether or not a long run equilibrium relationship between house prices in different regions of the UK actually exists, and therefore it is possible to create two sides to the argument. Most commonly authors have employed the Engle & Granger (1987) or Johansen (1988) ratio tests of cointegration to detect convergence. As mentioned previously the most significant areas of convergence are North and Scotland, according to the research of Cook (2003) which heavily disagrees with the earlier and well respected work from Drake (1995).
Cook(2005): "A consistent pattern of asymmetric adjustment is observed, with reversion to equilibrium occurring more rapidly (slowly) when house prices in the south east of England decrease (increase) relative to other regions" results support evidence of ripple effect whilst "the extent of cointegration uncovered cast doubt upon the recently proposed notion of weak segmentation".
The analysis of house prices at a national level has great history dating back to studies from Nellis & Longbottom (1981) and Hendry (1984), however recent years have seen the development of numerous studies on the relationship between house prices in different regions of the UK. Interest in these studies are due to reasons such as the distribution of wealth and labour market mobility in the economy (Alexander & Barrow 1994). Meen and Andrew (1998) also argued that an in-depth understanding of the housing market relies on the study of regional as opposed to national figures.
The possibility of a constant relationship between housing in different regions of the UK has been examined in works by Macdonald & Taylor (1993) Alexander & Barrow (1994) Drake (1995), Ashworth & Parker (1997), Meen (1999), Peterson et al (2002) and Cook (2005). Not all authors came to the same conclusion; however more recent studies have entailed similar arguments. When a long run relationship between regions holds, the ripple effect is supported. This is when changes in the house prices are firstly observed in the South East of the country before shifting to other regions.
Alexander & Barrow (1994) and Drake (1995) found minimal evidence to support the notion of convergence, whilst Macdonald & Taylor create the argument which has become known as weak segmentation due to the resulting relationships between the south regions and then regions elsewhere.
Cook(2005) noted that previous methods used to examine
The very clear view of "x" created a following interest by many authors.
Data Collection
The data has been sourced from Nationwide, and as mentioned before is split into 14 categories; 13 regions and the UK as a whole. Each region contains the average selling price of a house for each quarter of each year since the end of 1973. For simplicity the data for the last quarter of 1973 has been removed so as to start from the beginning of a whole year.
Data Analysis
