by Angus McIntosh and David Hutchings
There is a well-established record of rental values and capital values nowadays relating to the commercial and industrial property market stretching back over more than a decade. Similar data do not exist relating to land values. This article aims to analyse office land values over this period to see whether a theoretical approach to land value analysis can help in understanding investment performance.
The background
In 1984(*) McIntosh presented a paper which aimed to understand the relationship between property investment and hi-tech development. One of the principal ideas put forward was that of an equity-linked mortgage with the equity element being represented by the land, while the building (which has a finite life) was funded on a mortgage basis. It stressed the advantages of this route owing to the existence of generous tax allowances at that time. While the tax allowances (except in enterprise zones) are less generous today, the concept is still applicable.
In the equity-linked mortgage calculation the following might apply.
As we know, the property investment and development market does not normally rationalise a development situation in this way. The investment yield is arrived at based on the intuitive knowledge of the market in terms of both market valuations and investment sentiment.
It is no doubt apparent from this type of analysis that many of the more recent debt-equity funding arrangements recognise the difference between building costs (which are in effect met by debt finance) and the potential equity growth which rests with the location of the land upon which the development sits.
In the example given the total rental income is £113,080 pa. What happens if rental values move upwards? Building costs move upwards? Interest rates on the development costs change? Other costs increase such as VAT or other taxation? The planning consent restricts the development?
It is because of these variables and their wide interpretation that the Lands Tribunal is sceptical of using residual land value appraisals as evidence of actual land values. In particular, when a development will be under way for more than one to two years, such analysis can become wildly inaccurate which is why a cashflow approach in such situations is far more appropriate. The design of the building created may also influence the value of the site.
However, for the invester and/or developer, despite the Lands Tribunal’s edict, a decision has to be made whether to proceed with development. Answers must therefore be found to various questions such as: Are land values generally under or over priced? At what point should he/she become involved in new development/investment?
The problems of market data
Data availability and reliability are chief constraints in producing a whole range of property research and analysis. While much attention has been paid to compiling and analysing databases of rental and yield information, land values have been relatively overlooked. Indeed, residential and agricultural land value series are really the limit of what is regularly available over a sufficiently long time to allow adequate analysis.
Assembling an index or database of land values is fraught with problems, not least owing to the subjectivity of the market and the uniqueness of each deal done. If, however, an index could be produced which was capable of being disseminated to a regional or city level, then it would be a great help in understanding the dynamics of development activity and, second, in the identification of levels or areas of future opportunity. However, one should not forget the sentiment of the market or even the entrepreneurial determination of some developers and investors who believe they can outwit, market trends and wrest a market share from competitors despite a poor level of demand.
Land value research
The objective of this research was to produce an index of development land values for the office sector of the market, which, through varying the input factors, would be capable of showing movements at a national, regional and key city level.
Following on from this, and using output from the Healey & Baker office rental growth model, an attempt can be made to forecast DLVs for contrasting office markets.
The residual method of calculating development land values has been employed to produce an index, the input assumptions being those taken to be standard for development appraisals. Input data are derived from a variety of sources. Rental data are extracted from the H&B Prime rental index and future rents from the H&B office rental growth model. Construction cost data are taken from the BCIS survey. Forecasts of economic variables are taken from several econometric and forecasting sources.
The index is adapted for subnational markets by adjusting the input variables. A prime rental series specific to the submarket in question is introduced, alongside a corresponding list of development yields. Finally, a BCIS weighting is given to the building cost index to reflect conditions in the particular locale.
For the purposes of this article the City of London is analysed as the first distinct market and central Birmingham as the second.
The results
The results of this attempt to produce a series of residual land values are set out in the graphs which relate to an office development of 1,200 sq ft gross. These appear to be encouraging in that they display a high degree of variability and can also be justified in the context of trends evident in any one market.
Figure 1 shows prime office development land values as derived from the residual calculation time series, in cash terms. The dominance of the City of London is clearly shown, with land values eight times those seen nationally. The pattern for the three markets is broadly similar, with an upturn in rental growth in the early 1980s combining with a decline in building costs to produce prolonged growth in land values.
In London this was followed by a period of stagnation in spite of the low cost of borrowing. This was due to poor rental growth and yields increasing in 1982.
It was not until late in 1985 when rental growth began to accelerate in the run up to “Big Bang”, and interest rates peaked, that land values once more began to outgrow the national and provincial market averages. Prime development land values had grown from £1.5m in 1985 to nearly £4m in 1988 before escalating building costs and an unwillingness to pay yet higher rents blunted the upward surge.
In respect of Birmingham, the most striking feature is the negative values portrayed in the late 1970s and again in 1987. While negative land values are largely a theoretical idea, what can be deduced from this is that the city has offered a highly unattractive environment for office development over much of the past 11 years. When rental values are low and investment yields high, the gross development value may be insufficient to pay for the cost of construction, hence a negative land value.
Many provincial markets suffered a large oversupply of office space in the late 1970s, consequently reducing development opportunities. While rental growth accelerated in 1978 to ease this situation somewhat, rapidly advancing interest rates (from 5% in 1977 to 17% in 1979) halted the trend towards feasibility. However, this trough was short-lived and increasing rents, a fall in yields and the aforementioned decline in building costs produced strong growth. This resulted in positive development land values in Birmingham in 1980 and a period of stability lasting into 1984. The downturn in land values which followed was, however, more severe and longer-lived than for the London market owing to rising building costs at a time when rental values showed very little upward movement.
By 1987 the lack of rental performance in the Birmingham market had led to yields being large enough once again to create negative land values, despite interest rates being at their lowest level for a decade. However, the ripple effect of growth away from London reached Birmingham just at the time that the London market itself was slowing. A combination of falling yields and rapid rental growth brought about a more favourable investment climate than had been seen in Birmingham for more than a decade (though growth remained below the national average).
The progression of development land values demonstrated by this analysis illustrates the dynamism of the development equation as well as the thin line between feasibility for many provincial markets. The City of London is obviously a much stronger market, but the sheer level of land values can be a deterrent to development in itself. While these series may not exactly reflect actual development values, they do clearly indicate the feasibility of development at any one time.
Site value as a percentage of development value
Figure 2 indicates that portion of the gross development value (less site acquisition costs) which is attributable to the value of the site. Again it is highly noticeable that the responsiveness of each market varies markedly. However, it is also notable that the most volatile markets are the ones in which land values represent a relatively lower portion of the total development value. This may be a reflection of the continuity and stability engendered by high land prices or rather by the factors which cause them.
In the City of London, land has increased as a portion of total development value by some 0.6% per annum. In Birmingham, the growth has been of the order of 2.7% (owing to the large negative value at the period start) and at 1.25% nationally. In other words, an increasing amount of the value of any property investment is to be found in the land itself. Following this logic, one finds that the building is increasingly becoming an “expendable” part of the total investment value.
Forecasts of office development land values
Figure 3 depicts the index of development land values derived from the previously described time series of cash values. This graph clearly highlights the disparity in growth rates seen between different markets. London shows a much smoother and sustained growth path than either the national or provincial markets. Indeed, growth had been fairly consistent at a compound rate of about 17%. The marked volatility in other markets is clearly indicated, as is the rapid growth seen in 1979 and again in 1986-87. The current compound growth rates recorded for Birmingham and the national average are 15.8% and 21.6% respectively.
The forecast movement in land values in each market is also indicated in Figure 3. As with all forms of forecasting, it should be remembered that one is relying on the validity of input data (both historic and forecast) and also that the market will continue to operate in the future as it has in the past. Furthermore it should be borne in mind that, as with a valuation, the prediction is made at one point in time and its validity is constrained by the information then available, but may change as new factors become known.
It can be seen that the present dormancy in the City of London will continue into 1990. However, a tightening in the supply and demand balance and increased confidence in the market’s growth potential will bring about a return, later in the 1990s, to the order of land value rises seen in the period 1985 to 1987.
A different story is revealed for the national and Birmingham markets. Current high growth will persist into 1990, but may then reduce owing to a slowdown in rental growth and a tightening in the yield position. However, this should be short-lived and growth will return in the mid-1990s.
While the recent introduction of VAT to building construction has not been included in this analysis, the forecast could be adapted for a variety of VAT scenarios.
Conclusions
Although the analysis detailed above is as yet in its infancy, the results would appear to support the principle being carried forward.
(1) There has been an interesting difference in recent years between the investment performance of property generally and the change in land values. Land values have been theoretically far more volatile.
(2) There are clear structural differences between the two markets analysed in this article, which helps to provide a clearer picture of past performance of land values.
(3) In all markets the land value as a percentage of gross development value has been slowly rising. This implies that the problem of obsolescence is less important than has previously been appreciated, implying that buildings are increasingly expendable over the long term.
(4) As a method of forecasting future land value changes, the indices are a tool in helping to understand the feasibility of development and the timing of investment decisions.
(*) “Funding hi-tech industrial property”, EG 25 February 1984.
