T here are many aspects about property that still elude researchers, but probably none more so than depreciation. Yet, it is such a fundamental factor determining the long-term returns from the asset class. The first comprehensive study to attempt to quantify depreciation, ‘Depreciation in Commercial Property Markets’ was, using UK IPD data and CBRE data, undertaken in July 2005 (1) . Seven recognised previous studies had estimated the depreciation rate for office rental values, producing results ranging between 0.8% p.a. and 3.0% p.a. Some studies had also looked at yield or capital value depreciation, although there was evidently difficulties in differentiating between the effect of depreciation and the other factors that influence yield shifts. An initial analysis of the IPD UK database by the July 2005 study revealed that, over the period 1981 to 2003, the average age of the properties had fallen slightly, from 26.8 years to 26.1 years. That fall masked two things. First, in terms of the IPD Index time series, in most years the average age rose by more than the 12 months that had elapsed, suggesting that investors (who were mainly institutions) were selling more modern property and/or buying older property. The exceptions were seen during cyclical bouts of development, of which the most striking period was between 2000 and 2003, when the average age fell by one year at the All Property level. Second, there were significant differences between the sectors. Standard retail units’ average age rose by 13.9 years, industrials by 6.9 years, while offices fell by 0.3 years. Within the office sector, West End of London had become older, while the City of London and South East of England had become younger. The oldest properties were in standard retail units which, by the end of 2003, had an average age of 62.5 years. The fact that standard retail units had outperformed offices over the period and, yet, had the greatest ageing, raises an important question. What are we trying to measure in depreciation? The report fixes that firmly as meaning the loss in value due to age, although there is an acknowledgement that some locational depreciation may have been included in the numbers. But if depreciation equates to age, then it suggests that the investor universe, as measured by IPD, was either oblivious to it (by allowing the portfolio to depreciate and not countering it), or it believed that it was a factor less significant than the other ones that determine performance (such as sector, or building quality). There is also, however, the locational issue, to which the report made passing reference. Clearly when we are considering ageing, we are merely referring to the building: the land on which it sits does not age. Indeed, for as long as populations grow and their demands increase, the finite resource of land should see increasing competition for it. Where that demand is strongest – in recent times in city centres – we should expect appreciation to be strongest. And where the economic utility of the land is highest – for instance housing highly-paid office professionals – the land value should be highest. These locations will, to some extent, change over time, and that presents a particular problem in trying to estimate depreciation rates for the combined value of land and buildings. Nevertheless, the report results are shown in the table below The lower depreciation rate in the more recent period implies that the rate of inflation has slowed, but the report was not able to offer a reason for that. It should also be noted that there are a couple of negative figures in the table – although not very large – but the report could not fully account for those. Nevertheless, the 1% All Property figure, taken from the full period of data, has become something of ‘a rule of thumb’ amongst researchers and although it is derived from UK data, has received a universal application in developed European markets. There was, however, some concern in the research community that the sample used for the analysis understated the rate of depreciation because it was more prime than the universe of all properties. Of course, that pre-supposes that prime property depreciates at a faster rate than older secondary property.

Europe next and then back to the UK

An IPF report (2) published in March 2010 on depreciation in Europe used a similar methodology (3) to the UK analysis, but produced quite a strange set of results. Rental depreciation rates, over a 10-year period to 2007, ranged from 5% p.a. in Frankfurt to an appreciation rate of 2% p.a. in Stockholm. At the time, this report sparked much discussion seeking an explanation for the results, which included questioning the quality of the rental value data, although there was no wholly satisfactory conclusion. There were also many inconsistencies in terms of the age profile of depreciation, which appeared to vary between cities. It was suggested that depreciation seems to increase in stronger lettings markets so that the existing stock seem to lose out to newer properties at such times. However, when markets are weaker, existing properties do relatively better than new by not depreciating as much. I think that that was a particularly interesting observation, which I interpret as meaning that new developments (or refurbishments) are a causal effect of depreciation – to the existing stock. While there would still be physical depreciation to the existing stock without new developments, that can be largely fixed by maintenance expenditure. But if there is no better product, the effects of economic or functional depreciation would be much reduced, maybe to zero, and those are the more important forms of depreciation – rather than physical depreciation – in our evolving markets. A later report on the UK (4) used a different approach (5) , but this was applied to only offices and industrial property conceding, to some extent, the difficulties of applying it to retail property. An early question addressed by the report was the shape of depreciation over time. Is it linear, geometric (fast at the beginning), s-curve (slow, fast, then slow), or even a one-off event? While acknowledging that previous studies had produced no consensus, the report concluded that high quality properties, as measured by their rental values, suffered from higher rates of depreciation, and that depreciation rates seem to slow for very old properties. Nevertheless, the findings for age- related variations were disappointing, suggesting that it is quite difficult to identify the relationship, assuming that it exists. When academic studies fail to get to a set of fully satisfactory conclusions, there is always a temptation to rationalise the causes, such as the quality of the raw data. Of course, if the causes could be properly identified, then it is likely that they could be addressed in some way. My belief is obsolescence is too complex to model properly with the limited data that we have available. One only has to consider, as a starting point, the three different forms of obsolescence: 1 . Physical, defined as the loss in value due to ageing and wearing out of the building and its services 2 . Functional, which is the inability of a building to provide the economic utility required by the occupier in terms of the use for which it was built. This may be because the requirements or processes of the business have changed, often because of technological advances (such as the introduction of air conditioning or computers), although sometimes because standards (say regulatory) change 3 . Economic, also called locational obsolescence, the loss in utility due to factors external to the property itself. For example, the opening of a new shopping centre may move the ‘prime retail pitch’ to a different location

Too difficult

Add to this mix the two components of a property – the physical building and the land – which may experience depreciation (or, in the case of the land, appreciation) at different rates. Then there is the cycle which causes depreciation to cluster in the down-phases of markets. Each economic and property cycle has its unique characteristics in terms of drivers, which have different effects on depreciation rates. Put all of these factors together with buildings and tenants that each form a unique combination, and it is, I think possible to understand the difficulties of using statistical techniques to achieve a simple answer as to what is the rate of depreciation. But if it is too complex even for a mathematical analysis, does that mean that we cannot get a grip on this subject? I do not believe so, and my approach is simply this. All buildings were built for a purpose. The ones in the best locations at the time would have been built as prime. Consider one of those; an office building in the City of London built 20 years ago, for example. What is its rental value now? You might estimate that to be GBP50/sq ft/year. Now imagine that you were building on a prime location today, to a quality specification and modern requirements. What is the grade-A rent per square foot? Say GBP75/sq ft/year. The yields may now be 5.5% for the older building and 4.5% for the new one. A rough and ready calculation places capital values of, respectively, GBP900 and GBP1,650. The difference represents a measure of the depreciation of the building over the period. The answer is 2.0% p.a. for the rental value, and 3% p.a. for the capital value. Obviously, the answer will vary between sectors, locations and building types, but the approach does have the benefit of including all forms of depreciation. For Paris CBD, where the standard form of quality office building is a ‘Haussmannian’ five-story, plus basement and attic rooms, building constructed somewhere between 1853 and 1920, the rate of depreciation is very low – simply because there is little development to replace them. In effect, the planning restrictions have depressed depreciation within central Paris, and there is little difference in value between a building of 1850 and of 1920. While there is the ’other Paris CBD’ of La Défense, which offers relatively modern large floor plates, the differences between the two are so great, that the substitutability is very limited. Paris CBD depreciation is therefore more physical than economical. These are rough approaches to the issue but, I would argue, it is better to be roughly right than precisely wrong. Of course, you will then need to think about whether depreciation will be faster in the future – as I would argue – than it has been in the past. That is where understanding the drivers helps. Footnotes (1) ‘Depreciation in Commercial Property Markets’, V. Law, N. Crosby, S. Devaney, and A. Baum, sponsored by IPF, July 2005. What is known as cross-sectional regressions (2) Depreciation of office investment property in Europe, Neil Crosby, School of Real Estate and Planning, University of Reading, S Devaney, M Frodsham, R Graham, and C Murray, March 2010, funded by the IPF Research Programme published in March 2010 on depreciation in Europe used a similar methodology (3) This was what is known as a longitudinal study, but the basic methodology was fairly simple. The IPD data represented the performance of real property, including depreciation. The CBRE data – like much other agents’ data – represented hypothetical prime property, which is assumed to be prime at every data point and, therefore, ignored depreciation. Subtracting one from the other should, theoretically, produce a rate for depreciation. (4) Modelling Causes of Rental Depreciation for UK Office and Industrial Properties, part of the IPF Research Programme, N Crosby, S Devaney, and A Nanda, June 2013 (5) What is known as cross-sectional regressions

Ltd