Economics & politics Research Briefing May, 11 Europe s office markets: Rent cycles weakening In many European markets for office space vacancy rates are now higher than at earlier turning points. This structural change has farreaching consequences, also regarding adjustment to market cycles: we can show that higher vacancy rates reduce the volatility of office rents. For Europe s current economic recovery this means that the potential for rent increases is lower than in the 199s. There are only a few office markets to which this general rule does not apply. It is especially true, though, for office markets in Germany and applies not only to the large A-rated cities but also to the smaller B-rated and C-rated cities. There, the relative value of investing in B-rated locations to avoid market volatility has declined somewhat. For opportunistic investors, timing is important above all in the large office markets. If speculative construction activity remains low, it is by no means assured that this rule will continue to apply well beyond the next economic cycle, as the ageing stock of office space will provide some supply-side relief. Authors Tobias Just +9 69 91-31876 tobias.just@db.com Jaroslaw Morawski RREEF Research +9 69 717-6 jaroslaw.morawski@rreef.com Editor Bernhard Speyer Technical Assistant Sabine Kaiser Deutsche Bank Research am Main Germany Internet: www.dbresearch.com E-mail: marketing.dbr@db.com Fax: +9 69 91-31877 Managing Director Thomas Mayer Large and small hog cycles Demand for office space depends heavily on the region s overall economic development. In an upswing phase, jobs are created faster than new office space. Vacancies are reduced and rent prices rise. This in turn attracts property investors: the additional supply of space often enters the market with a time lag and in some cases at such a late stage that demand for office has already begun to decline again. Especially as adjustment processes in property markets are very long and drawn-out, pronounced hog cycles may emerge and additional supply hits the markets at the worst possible moment. In the downswing phases vacancy rates will then rise noticeably above their long-term average, depressing rent prices for office space. This pattern has been observed in many European office markets over the last few decades, with sometimes considerable regional differences especially during the financial and economic crisis: in markets such as Barcelona and Dublin, where a particularly severe recession followed a long upswing phase, supply growth was still geared to the strong pre-crisis years which then led to particularly pronounced hog cycles also compared with earlier recessions. In other markets such as or Amsterdam where the legacy of the previous office market recession was a drag on growth even during the last upswing, the financial and economic crisis resulted in a markedly weaker hog cycle: office rents in and Amsterdam only rose moderately up until 8 and declined only slightly in the following years.
Research Briefing Vacancies and prime rents are related x-axis: Vacancy rate in (%) y-axis: Prime rent in (EUR) y = -.53x + 3. R² =.33 5 1 15 Q1 1991 to Q 1 55 5 5 35 3 5 Sources: CBRE, DB Research 1 : From low to higher vacancy rates x-axis: Vacancy rate in (%) y-axis: Prime rent in (EUR) y = -.x + 5.5 R² =.89 y = -.5x + 3.6 R² =.1 5 1 15 Q1 1991 to Q Q3 3 to Q 1 55 5 5 35 3 5 Sources: CBRE, DB Research Rental ranges in Munich x-axis: Vacancy rates (%) y-axis: Prime office rents (EUR/m²) Rental range (199s) EUR 1. Rental range (s) EUR 3. 6 8 1 1 Q1 1991 to Q Q3 to Q 1 38 36 3 3 3 8 6 Sources: CBRE, DB Research 3 In this paper, our primary objective is not to show that the adjustment paths of European office markets differ from region to region, but rather that these adjustment paths are not stable over time. This fact is very important for the forecasting of key office market figures since formal forecasts, based for instance on econometric models, rely on historic data. They extrapolate past relationships into the future. Of course this is a problem not only for formal forecasts. In the end analysis, the heuristics used by practitioners are a reflection of long-term experience considered to be stable. Creeping structural change or even hefty structural breaks in office markets can trigger changes in market processes. If one relied on the usual yardsticks such as estimated coefficients, one would make systematic errors. The office market is a case in point (see charts 1 and ): plotting the vacancy rate for against that city's prime rent in the period 1991-1 yields the usual negative correlation the higher the vacancy rate, the lower the prime rent. Each additional percentage point added to the vacancy rate will result in an EUR.5 decline in the prime rent. This is plausible but only half the story. For the office market the millennium marked an important turning point. The supply of office space increased strongly in the course of the dotcom bubble. The trend line depicted in Chart 1 with a slope of.53 is the average of two very different phases. This is illustrated in Chart : from 1991 to the relationship between the vacancy rate and the prime rent is depicted by a much steeper line than in the years following. In the 199s vacancy rates varied between.5% and 1%. During that time of relatively low vacancy rates, office space quickly became scarce. Each 1 percentage point decline in the vacancy rate drove up the prime rent by EUR.. After the average vacancy rate was between 1% and % and prime rents then reacted considerably less to changes in the vacancy rate (EUR.5 per percentage point change). Less rent volatility not only in am Main A similar phenomenon was also registered in Munich's office market. Over the past ten years, the vacancy rate in Munich was markedly higher than in the 199s, with rent volatility decreasing strongly over the same period. For purposes of illustration we determined the rental range by calculating the difference between the highest and lowest prime rent in a certain location for a given period of time. In Munich this range stood at EUR 1. in the 199s and fell to a mere EUR 3 after. In, the range also shrank by one-third. In Hamburg, which had not seen a similar disruption between the 199s and the period following the dotcom crash, the rental range has been only marginally smaller after than in the 199s. 1 The transition from lower to (markedly) higher vacancy rates also led to a situation where the pace of adjustment in these three German markets was generally slower. This has far-reaching 1 There is an interesting side note provided by these calculations. Most market participants have got used to calling the most volatile market in Germany. If we look at the relative rental range as a measure of volatility, however, this is no longer so easy to affirm based on the last ten years. The fluctuations in were only higher than in Hamburg of late because the absolute rent level is higher, while the relative changes are similar to those in Hamburg. A simple linear regression of peak rent against the vacancy rate (see chart for ) shows that for the period to 1 both Munich and Hamburg actually boast higher coefficients than. A change in the vacancy rate by one percentage point currently increases/reduces the prime rent in Munich and Hamburg by an even larger amount than in. May, 11
Europe's office markets: Rent cycles weakening implications for both investors and forecasters as simple heuristics or model estimates based on very long data series may have to be recalibrated. Vacancy rates trending up Vacancy rates for European office markets, 1-year moving averages 199 1995 5 1 198-1 sample 1985-1 sample 199-1 sample Sample 1 included 6 markets, sample covered markets and sample 3 included 8. Sources: PMA, RREEF Research 1 1 8 6 Less volatility in three German office markets (1991-) (-1) Munich (1991-) Munich (-1) Hamburg (1991-) Hamburg (-1) Mean rent (1) Mean vacancy rate Rental range EUR/m² % Absolute () % (3) 39.1 5.7. 5.1 36.3 15.9 6.5 17.9 3.5. 1. 3.1 3. 8. 3. 9.9.8. 6...3 8.. 17.9 (1) The mean rent is the average prime rent in the respective period. () The absolute rental range is the difference between the respective highest and lowest prime rent in a city during his period. (3) The relative rental range is the ratio of absolute rental range to mean rent. Sources: CBRE, DB Research 5 Europe's office rents less volatile Standard deviation in peak rents, 1-year moving averages.35 199 1993 1996 1999 5 8 198-1 sample 1985-1 sample 199-1 sample Sample 1 included 6 markets, sample covered markets and sample 3 included 8..3.5..15.1.5. Sources: PMA, RREEF Research 6 Europe s major office markets The trend witnessed over the last years of a gradual, in some cities even sudden rise in vacancy rates and a concomitant decline in the volatility of office rents is not restricted to the German office markets. To demonstrate this development we started by calculating moving averages for vacancy rates as well as the standard deviations in prime rents for several, variously sized random samples of European office markets. These samples were defined solely on the basis of data availability. The moving averages were calculated over 1 years in order to largely exclude cyclical influences. We examine three different periods of time, the first from 198 to 1. Due to the moving average, the first data point is the year 199. This sample comprises six markets. The second sample covers the years 1985 to 1, comprising no fewer than markets. And finally the third, very short sample, covers the period 199 to 1 for 8 markets. Chart shows that in all three periods the moving averages of vacancy rates for all markets in the sample tend upwards. There is, consequently, a very strong trend towards high vacancy rates in Europe. At the same time, Chart 5 shows that the moving averages of the standard deviation for the respective prime rents have declined. 3 In all three samples, the markets have on average 3 It is always the unweighted means that are depicted. We have also carried out calculations using weighted means, but the differences were extremely small on account of the strong correlation between the individual markets. Without the smoothing of the variables the decline in volatility in the rental markets is also shown (of course) in reduced short-term adjustment reactions. For two of the three surveyed European cities the standard deviations in the annual rental May, 11 3
Stuttgart Cologne Edinburgh Rotterdam Amsterdam Manchester Munich Rome Dusseldorf Birmingham Berlin Brussels Dublin Hamburg Glasgow Madrid Lyon Copenhagen Helsinki Vienna London: City Barcelona London: West End Oslo London Docklands Research Briefing become less volatile in both directions, upwards and downwards. Also, correlation analysis reveals that there are very few exceptions to this rule (see chart 7). Only a few exceptions to the rule Correlation between vacancy rate and rental volatility (1-year moving averages) 1..5. -.5-1. Values under zero signify that there is a negative correlation for this city. Values higher than zero signify that there is a positive correlation. Sources: PMA, RREEF Research 7 Lower volatility in Europe's office market x-axis: Rise in vacancy rates (percentage points) y-axis: relative rental range (%) y = -1.x - 1. R² =.17 - - -6-8 -1-5 5 1 15 Country group with average volatility Country group with very high volatility Sources: CBRE, DB Research 8 Pronounced differences remain between individual markets Next, we look at individual developments in Europe s office markets. To do this we again compare two periods, from 1991 to 1, on the one hand, and from to 1 on the other. Out of all the 33 office markets 5 analysed there were only five in which the average vacancy rate was lower during the second period than in the first (Milan, Paris, Budapest, Glasgow and Manchester); the respective decline in the vacancy rate was modest in all five cases (maximum 1.5 percentage points). For the other 8 markets the vacancy rate rose by nearly percentage points on average. In a few cases the increases came to 1 percentage points ( am Main, Dublin). This trend towards higher vacancy rates has reduced the volatility of office rents in very many European markets. In 31 of the 33 markets analysed the relative rental range 6 after has been smaller than in the ten years until 1. Across all markets the range narrowed by nearly 3 percentage points and in a few cases by even much more than 7 percentage points (e.g. Milan, Stockholm). Only in the cases of Oslo and Birmingham did the city sample reveal a widening of the relative rental range. If we then plot the change in the vacancy rate against the relative rental range for the two periods we find a remarkable dichotomy within the group as whole: in the first and much larger group the drop in the rental range is only moderate and there is a significant relationship between higher vacancy levels and lower rent volatility. This is completely in line with the above-mentioned analysis of 5 6 changes in the years after were a long way below some of those in 1991 and 1. For the sample from 199 to 1 these were four out of 8 markets, namely Barcelona, Oslo and sub-markets of London. For several cities there was data for sub-markets. We calculated the relative rental range as follows. Firstly, the difference was calculated between the highest and lowest rent levels during the period, and this absolute rental range was then set against the mean rent for the period. A value of 5% indicates for example that the rent for a location with a mean prime rent of EUR 5 could fluctuate between EUR 37.5 and EUR 6.5. May, 11
Europe's office markets: Rent cycles weakening Fewer ups and downs Relative rental range (%) Dublin Stockholm Madrid Milan Prague Barcelona Berlin Lisbon Helsinki London Amsterdam Oslo Warsaw Paris Zurich Budapest Copenhagen Brussels 6 8 1 1 1991-1 -1 The relative rental range is the difference between the highest and lowest rent relative to the mean rent for the respective period. Sources: PMA, RREEF, DB Research More and more vacancies Vacancy rates in German cities (%), moving averages 1999 1 3 5 7 9 A-rated cities C-rated cities B-rated cities 9 1 Sources: BulwienGesa, RREEF Research 1 9 8 7 6 5 3 Europe s office markets. The second and very small group is highly heterogeneous and comprises markets with a reduced rent volatility of more than 5 percentage points. Although all the cities in this group also registered a major rise in their respective vacancy rates a simple linear relationship could not be derived. Special structural factors appear to have played a major part in their case. This second heterogeneous group includes cities such as Berlin and Prague for which the economic transformation process was pivotal in the199s. Dublin, Barcelona and Lisbon in the second group of countries were also in a special structural situation on account of monetary union (or other local factors) and the accompanying interest rate convergence in the 199s. For all these markets at least a larger share of the reduced volatility was due more to the end of special factors than the simple transition from low to higher vacancy rates. This effect was a supplementary factor for cities in the second country group. Germany s larger and smaller office markets In the first section of this paper we have shown that for example for the cities of Munich and the marked increase in office vacancies led to much lower volatility in office rents over the last ten years than in the 199s. Munich and am Main were those markets that initially benefited most from the euphoria surrounding the Neuer Markt and as a result experienced the largest expansion in office space in the following years. Furthermore, we show that the phenomenon of declining fluctuation in office rents is by no means exclusive to the few prime locations in Germany. The decline is merely most pronounced there. Adopting a similar approach to office markets in Germany as to those in major European cities, we have calculated moving 1-year average vacancy rates as well as the standard deviations of the peak rents and mean rents. We have carried this out for each of the three groups of cities which have been put into categories A, B and C. 7 The A-rated cities have seen the biggest rise in vacancy rates over the last few years: the moving average for the seven biggest German office markets has risen over the last 1 years by 5 percentage points, for the B-rated cities by about 3 percentage points and the C-rated cities by.3 percentage points, for the smaller cities the increase has slowed markedly in recent years. For all location categories the volatility (again the respective moving average of the standard deviations) of peak rents fell considerably. The difference in these standard deviations for the three market categories (A, B, C) has barely changed in the process. The fluctuation range for the A cities, though, is barely higher than for the smaller locations in 1999. Also the volatility of average rents trended downwards as vacancy rates rose. There is one striking aspect though that perhaps becomes most apparent if one looks at how the difference between prime and average rents has developed: this difference measures the premium that has to be paid in a city for the very best properties in prime locations (CBDs). In A-rated cities this premium fluctuates by 5% with the office market cycle. During an upswing prime 7 Market size is very important for the categorisation, even though it is not the sole determining factor. The group of A-rated cities comprises the seven most liquid office markets in Germany (Berlin,, Dusseldorf, Munich, Hamburg, Cologne and Stuttgart). The B-rated locations are the 1 next biggest cities and the C-rated city group is made up of the next largest cities. We also analysed D-class cities. We confine our comments to the results for the first three groups, however, since these are more important for institutional investors. May, 11 5
Research Briefing Less fluctuation Standard deviation in prime rents of German cities (%), moving averages 1999 1 3 5 7 9 A-cities B-cities C-cities Sources: BulwienGesa, RREEF Research 11 Rent premia for prime locations Prime rents relative to mean rents in German cities (%) 199 1995 5 1 A-cities B-cities C-cities 6 5 3 1 Sources: BulwienGesa, RREEF, DB Research 1 3 1 locations do better than the average and they are also harder hit by recessions than peripheral locations. Across all cycles this premium remained constant, however. 8 In the smaller B and C cities the rent premia for prime sites have by contrast risen since the mid-199s by and large. Admittedly this was due less to the especially strong rent increases at the peak (volatility had dropped sharply particularly in these cities) than the very low to negative growth in peripheral locations. 9 This means that in A-rated cities the fluctuation narrowed by a similar amount in both the centres and the subsidiary locations. The prospects of rent increases in these cities during an upswing are, however, still much greater than in secondary locations. The same applies to the downside risks in years when the cyclical downturn is underway. In the smaller B-rated and C-rated cities the fluctuation range has converged very heavily. Timing and above all the concrete location appear to be not only less important for these smaller cities than for investments in A-rated cities, these two factors are currently also much less important than 1 or 15 years ago. Three main conclusions can be drawn from these considerations for property investors in Germany: The current economic upturn will drive up demand in most office markets. The potential for rent increases associated with reduced vacancies will, however, probably be much lower than during the upturns in the 199s in the same way as during the upturn until 8. Overall, the decline in volatility of German real estate is interesting especially for all those long-term investors for whom the stability of the cashflow is more important than timing-focused value-changing potential. The important thing here is that the rental yield already correspondingly offsets the increased risk of default resulting from the higher vacancy rate. Inside Germany there are fewer bargains to be had by opportunistic investors than in the 199s at least in terms of the probability of benefiting from cycles. Timing and location are much more important in A-rated cities than in the smaller cities. Regardless of this, the German commercial property market is more attractive than the southern European markets on account of its very specific interest rate environment. Conclusions In very many European cities the structural increase in vacancies over the last ten years led to declining volatility in prime and average rents. The potential for rents to rise in many European cities is therefore likely to be lower during this upswing than in the 199s. Two important qualifications nevertheless need to be made in closing: 8 9 The process of calculating the mean values for the location groups hides the interesting distinctive regional characteristics. The highest premia for downtown locations are paid in am Main, while in Munich the premium is trending up most sharply, whereas in Stuttgart the premium has fallen considerably. The mean values are only indicative, there are no laws that suddenly change at the boundaries of the groups. Interestingly in the B-rated and C-rated cities there was a much stronger convergence between the rent volatilities in central and peripheral locations than in the even smaller D-class cities. It may be that for the very small cities the structural effect of rising vacancies is amplified by the structural effect of the population concentration in major centres and that this weighs most heavily on the outskirts of small towns. 6 May, 11
Europe's office markets: Rent cycles weakening Regional differences remain grave despite this very widespread trend: given the severe downturn endured rent volatility in Ireland and Spain was even more pronounced than during the downturns in the 199s. The development described above should not simply be extrapolated into the future. For instance, speculative construction projects were already less of a factor during the boom years until 8 than in the late 199s. On the supply side a key reason for the structural increase in vacancies has thus been mitigated. The ageing of the office stock is a delayed function of past construction activity. In particular the technical and energy efficiency of buildings are increasingly limiting the rentability of very old office property. The phases of very large increases in supply in the 198s and 199s will thus exert knock-on effects on the volume of replacement investment. This will probably not yet be a relevant factor during this upturn, but this could already be the case in the next cycle. The result could then be a next phase of structurally declining (market-moving) vacancies and correspondingly higher volatility in rents. Tobias Just (+9 69 91-31876, tobias.just@db.com) Jaroslaw Morawski (+9 69 717-6, jaroslaw.morawski@rreef.com) Copyright 11. Deutsche Bank AG, DB Research, D-66 am Main, Germany. All rights reserved. When quoting please cite Deutsche Bank Research. The above information does not constitute the provision of investment, legal or tax advice. Any views expressed reflect the current views of the author, which do not necessarily correspond to the opinions of Deutsche Bank AG or its affiliates. Opinions expressed may change without notice. Opinions expressed may differ from views set out in other documents, including research, published by Deutsche Bank. The above information is provided for informational purposes only and without any obligation, whether contractual or otherwise. No warranty or representation is made as to the correctness, completeness and accuracy of the information given or the assessments made. In Germany this information is approved and/or communicated by Deutsche Bank AG, authorised by Bundesanstalt für Finanzdienstleistungsaufsicht. In the United Kingdom this information is approved and/or communicated by Deutsche Bank AG London, a member of the London Stock Exchange regulated by the Financial Services Authority for the conduct of investment business in the UK. This information is distributed in Hong Kong by Deutsche Bank AG, Hong Kong Branch, in Korea by Deutsche Securities Korea Co. and in Singapore by Deutsche Bank AG, Singapore Branch. In Japan this information is approved and/or distributed by Deutsche Securities Limited, Tokyo Branch. In Australia, retail clients should obtain a copy of a Product Disclosure Statement (PDS) relating to any financial product referred to in this report and consider the PDS before making any decision about whether to acquire the product. May, 11 7