BRAZIL Cushman&Wakefield Research Department AUGUST 2015 THE TRICKY PRICING OF REAL ESTATE Price is the key driver in economic life. It measures the value of one good against all others, juggling supply and demand into market equilibrium. Yet, this measurement is not always easy. The more market characteristics are closer to a monopoly (in contrast to perfect competition), the less prices are free to regulate themselves. Valuations are a good example of this. Building owners and investors demand valuations because they lack the market knowledge and expertise to price their properties, and they need to support their asset values with solid, rational assumptions. Rents move in much the same way, but in another vein. Space for tenants is a necessary business investment in order to achieve high productivity, which, in turn, must match its business yields. Call centers, for instance, will never occupy the most desired real estate location. Those spaces are typically reserved for more profitable sectors such as financial or high-end industry corporate employees. In this sense, finding the best space for each business type can prove to be tricky, and help is often needed. This is where market experts come in: to supply market knowledge and expertise, and to help to find for what is best for their clients. In both cases, market imperfections (buildings are not homogenous products, their supply is limited, information is not available to all market players etc.) make expert intervention necessary to decrease risks, and consolidate market rents and yields. Cushman & Wakefield South America Praça Prof. José Lannes, 40 4º andar. São Paulo, SP. Cep: 04571-100 - Brazil www.cushmanandwakefield.com AUGUST 2015 1
GETTING TO THE BOTTOM OF RENTS There are two types of rents: asking and effective. Asking rents correlate with available spaces. Once a space is available for lease, the building owner asks for a rent that will meet his or her expected investment yields (or more). In most markets, prospective tenants then start biding, hoping to negotiate a discount. In a more tenant favorable market cycle, landlords compete for tenants, often dropping asking rents even further. Effective rents correlate with leased spaces. After the tenant and landlord negotiate, an effective price is met, and lasts for as long as the contract is held or at least for the first three years, after which rent values can be renegotiated. The effective rent is almost always lower than asking rents due to negotiations that typically include free rent, tenant improvements and builtouts. These two types of real estate rents can be used to calculate two different average market figures, the average asking rent and the average effective rent, each with a distinct dynamic. One cause of this different behavior is the market share in which each is calculated: the average asking rent is based on available spaces, commonly representing less than 20% of market rentable area, while the average effective rent is calculated with the remaining occupied space (%). Because of the smaller sample size, the first average is usually more volatile than the second. Another cause is that, in any time period, the asking rent of any availability can change in a more tenant (or landlord) favorable market cycle, whereas effective rents, with the exception of those under renewal or new leases, are consistent (or adjusted for inflation), which ensures a much more flattening trend. This would not be the case if we considered only new leases in the effective rent sample, but this is rarely the case. Having complete (100%) market information, though, almost never happens in the real world. Some markets are less transparent than others, which makes it even harder to collect and aggregate any figures. In the Brazilian market, for instance, rental contracts are not available to the public, and, even worse, they are often protected by non-disclosure agreements between landlords and tenants. That is why even companies like Cushman & Wakefield, that have access to a large number of contracts, have a very difficult time building an effective rent market trend. Moreover, searching and compiling non-disclosed information of at least % of the real estate market can be too costly. Asking Rent The price that theoretically meets the building owner expected investment yields. Effective Rent The effective rent is most of the times lower than asking prices due to negotiations between tenant and the landlord. This is why most of the real estate research in the country is based on asking rents. In summary, asking rents are (1) always available to confirm, (2) they are public information, (3) there are no non-disclosure issues, (4) they make a good market gauge as they are constantly changing, and (5) one only needs to compile information of about 20% of the market to calculate an average. Although this is easier, it s still a challenge, and it gets more complicated in a market where landlords do not use one common type of area measurement. Even though contracts are obliged to be based on a government regulated area type, much of the asking rent information that landlords release to the market do not follow these rules: they often use larger rentable areas, promoting the building s common spaces, which decreases asking rents per square meter. This obviously makes research work much harder, as it offsets true knowledge about the market. Average Asking Rent, rather than Average Effective Rent, is the go to aggregate for rents because: 1. It relies on public information 2. It is always available to confirm; 3. There are no non-disclosure issues 4. It has good Market gauge, as they are constantly changing 5. Only 20% of the market information is needed to calculate its average. 2
STICKING TO WHAT YOU HAVE PRICE INDEX, WHAT ABOUT THAT? At this point, it is safe to say that we have to use asking rents as the basis of any aggregate price trend that we might want to build. Now we have to deal with the downsides. For example, the average asking rent is more volatile the smaller the market inventory and/or vacancy levels are, because newly put availabilities in the period tend to have bigger impacts in the final average. In addition, for Brazil, asking rents might not use the same area type, forfeiting the figure levels. However, as worrisome this might seem, there are still ways to make good use of this information. One way is not to look at the levels, but to the overall trends instead. Although the average asking rent level might be overestimated (or underestimated), its quarter over quarter (QoQ) or year over year (YoY) change should be right, since landlords rarely change the way they estimate their rentable areas. Still, some important issues need to be addressed. For one, it is only when the market cycle is near its peak (either tenant or landlord favorable) that asking rent averages show a clear trend. If the market is not near its peak, the average asking price will be skewed by volatility. One way to overcome this is to look to at yearly data, but that has the disadvantage of not being able to display short-run trends, which are in a broker s best interest. Moreover, using average asking rent trends for modeling and forecasting the real estate market can be misleading as the average might rise with newly available high priced availabilities put in the period, and fail to take into account the fact that individual asking rents are falling on a quarterly basis. This makes it impossible for the model to properly correlate actual asking rent changes to demand for space, meaning that econometric regression coefficients (e.g. asking rent elasticity) would be wrong, as price variations are not accurate for the model to measure its sensibility against demand. And finally, if one is to compare the average variation to any public price index (in Brazil, we use IGP-M for lease contracts and IPCA as our Consumer Price Index), the comparison will not be entirely correct, since the average asking rent variations are not rent variation isolated, but rather a mix of rents and availability changes between periods. Everyone in the industry is familiar with price indexes. The most common is the Consumer Price Index, or CPI; (in Brazil it is called Índice de Preços ao Consumidor Amplo - IPCA), which measures how the average cost of living is increasing (or decreasing) due to the rise (or fall) of prices over time. More commonly referred to as inflation, it is the key element behind labor disputes, any Central Bank inflation targeted work, and real asset yields. Of course, few people know how to technically perform its calculation, but all get the idea that it is measuring price change at an aggregate level. That is exactly what we intend to do with our real estate historical data. Adapting the price index methodology, we mean to isolate the rent change effect from the aggregate average asking rent trend. This would give us figures solid enough to access all problems discussed above: short and long run analyses, modeling, and index comparisons. We started with the raw data, aggregated it to calculate Paasche and Laspeyres indexes and, finally, derived the Fisher price index. Traditionally, the Laspeyers Price Index is the one used (i.e. CPI), because basket changes are much more difficult to be tracked. For instance, in Brazil, the consumption basket update for the IPCA occurs when the Instituto Brasileiro de Geografia e Estatística (IBGE) runs a survey called Pesquisa de Orçamentos Familiares (POF Research on Family Expense Budget), which should take place every 5 years the last one took place in the 8/9 period. In our case, we happen to know each period s basket of availabilities, so we were able to run all three price indexes. Statistically, Fisher is always the best out of those three. That said, we calculated our Cushman & Wakefield s Price Index (CWPI) with our data from São Paulo s office market. Excluding the asking rent price index influence from the average asking rent, we were able to calculate the effect that changing availabilities had in each period, or more specifically, our Cushman & Wakefield s Composition Index (CWCI). We discuss our findings below. As one may now conclude, the way you aggregate data matters. Regarding average asking rents, all issues have a close connection to newly put availabilities in the period. Our goal, therefore, is to exclude that effect from the average price trend, without compromising the actual asking rent behavior in the process. We, at Cushman & Wakefield, have developed a methodology to do so and our findings are outlined in the next section. Avg Ask Rent% = CWPI% + CWCI% 3
TENANT FAVORABLE LANDLORD FAVORABLE DECONSTRUCTING AVERAGE ASKING RENT DECOMPOSITION ANALYSIS In the last ten years, the São Paulo office market passed from a strong landlord market cycle to a more tenant favorable one. Availabilities, which were hard to find and much scrambled over before, are now giving tenants bargaining power over negotiations. Average asking rents increased significantly in this scenario, jumping 130% during the 6-2012 period, and falling -13% between 2012 and the last quarter of 2014. For the CBD market, these trends were even harsher: 206% in for the first period (6-2012) and an -18% decrease in the second period (2012-2014). Until 2012, the increasing asking rent trend was often used as a drive to motivate the construction industry for investing in the sector. This, plus the nurturing macroeconomic scenario, made building deliveries from 2012 until now sharply increase market inventory: they represent roughly 20% and 44% of today s total market and Class A rentable area, respectively. São Paulo: Market Cycles Slow Growth São Paulo 7 São Paulo 2014 Downturn São Paulo 2010 São Paulo 2012 São Paulo 2013 Accelerating Recovering Rent still elevated but falling from top of market cycle Falling rents promise future opportunity for tenants Rent at or near bottom of market cycle. Ideal for tenants leasing or seeking to lease property Rent growth accelerating. Ideal for owners of property Rent growth slowing. Still landlord favourable but growth is down from peak One question still left unanswered, though, is whether the average asking rent rise discussed above is based on increasing asking rents or on new and more expensive availabilities in the market. In other words, we intend to know exactly the impact of each in the average asking rent trend. The result we found, not surprisingly, is that both pushed asking rents up during the 6-2012 period. In other words, higher asking rents and more expensive new availabilities contributed to the average asking rent drift. Yet, curiously, most of it did not come from the first, but from the second. CWPI showed only a 31% increase between 6 and 2012, averaging 3.9% per year, while the CWCI showed a 76% increase, averaging 8.4% per year. This means that the average asking rent rise in that period had little to do with actual asking rent increase! Because most of the new inventory delivered has high specifications and come at such large rentable areas, the main source of average asking rent increase in our historical tables comes from the CWCI, not the CWPI. Interestingly enough, in the period from 2013 to 2014, the average asking rent fell by -10%, and was mostly influenced by asking rents (-13%) rather than new deliveries (+4%). At this point, market vacancy was already well above 15%, especially for Class A buildings (closer to 20%). Not only that, since a whole set of highend more expensive buildings were already in the market, recent new deliveries have had a smaller impact on the market s average asking rent in the last two years. São Paulo: Avg Ask Rent SP X CWPI SP x CWCI SP CBD + NCBD Class ABC (Index 6 = 100) 260 230 Avg Ask Rent SP CWPI SP CWCI SP Knowing that asking rents did not contribute as much as one could have expected, though, is still not enough. In Brazil, where the Central Bank s inflation target measured by the IPCA is 4.5%, a 1.6% CWPI yearly average increase from 6-2014 seems rather low. That is even more concerning when looking at the IGP-M, which commonly shows bigger figures than the IPCA. Because effective rents are yearly escalated by the IGP-M, this suggests there should be a smaller gap between asking and effective rents during the period. However, savings may have occurred on contract renewals or tenant relocation, compensating this trend. Either way, without the average effective rent data, it is impossible to come to a conclusion on whether one or the other effect prevailed. What we can do is to compare our CWPI with other indexes in the economy and that is the subject of the following section. 4
CWPI COMPARATIVES As expected, our CWPI for São Paulo s overall market ran well below the IGP-M and the IPCA for most of the period from 6 to 2014, especially in recent years. During this time, actual asking rents increased 14% (1.6% on average per year) while the IPCA and the IGP-M cumulated 55% and 66%, respectively, averaging 5.7% and 6.1% on yearly basis. Even though there is no rule that should connect the three of them, the difference is quite impressive. As it seems, the real estate market cycle is the main drive for the increase or decrease of rents, rather than overall price changes. Further, from the data one can see that vacancy levels, rather than vacancy changes, impact asking rents. Prices soared when vacancy levels where very low, and have fallen ever since vacancy levels got very high. Effectively, it is the difference between actual vacancy and its equilibrium level that impacts rents. In other words, after the amount of available spaces in the market reached a certain level, the negotiation environment began to change, impacting significantly asking rents. From 8 to 2014, both housing indexes had much greater variations: the Fipe-Zap index cumulated a 67% increase in the period, averaging 8.9% per year, while the IPCA sub item showed a 69% increase, or 9.2% per year. This is not new, as we saw before our CWPI performed well below the total IPCA and IGP-M indexes. If we look at the actual yearly variations, though, one is able to see the same deceleration seen in the CWPI as in the Fipe-Zap index, but not in the IPCA housing rent index. Now, there are two possible reasons for this. First and most obvious, the Fipe-Zap Index is based on São Paulo s real estate market. This means that the trend seen in our CWPI and the Fipe-Zap index share the same background some are the same developers, for instance. In this sense, the sources of absorption and investments are linked, as are new deliveries. Secondly, the Fipe-Zap index uses asking rents in its calculation, rather than effective rents. If what we discussed in the sections above is correct, more specifically, then overall effective rents tend to be more resilient than asking rents, and this could be the evidence needed to confirm it. CWPI, IPCA and IGPM (Index 6 = 100) CWPI, IPCA Rent and FIPE-ZAP SP (Index 6 = 100) 1 150 120 90 CWPI SP IPCA IGPM CWPI SP IPCA Rents Fipe Zap SP The next step is to compare the CWPI not to the general index, but to the best proxy we can find. In Brazil, we have two separate time series that track asking and effective rent for housing, including multifamily real estate: São Paulo s Fipe-Zap asking rent index and Brazil s IPCA sub item called residential rents. These markets certainly behave differently than commercial real estate property, but they are the best publicly available proxies in the country. 5
CWPI COMPARATIVES And finally, with regard to costs, we compared our CWPI to the indexes that measure construction labor costs and building material from the IPCA and IGP-M. Both of them suggest that these costs may be overburdening commercial real estate properties during the 6 to 2014 period, and display a much more significant increase than the CWPI. Construction costs rose the most, showing a 107% (9.5% year average) increase, and 96% (8.8% year average) for the IPCA and IGP-M subitems, respectively, against 37% (4.1% year average) and 57% (5.8% year average) increases shown in the building material sub-items. The latter should be more important as a cost source for building maintenance than the former, however, they both surpass our CWPI. SHEDDING LIGHT ON THE SUBJECT Our findings show that excluding the changing availabilities effect from the average asking rent calculation is a necessary step to better understand market rent trends. In our example, new high-end building deliveries had the biggest share in the last asking rent boom that took place between 6 and 2012, whereas asking rents per se did not change as much. Moreover, we saw that our CWPI cumulated changes that were well below any other indexes proposed, signaling value loss in the office market. It is important to note that any correlation between asking and effective rents made so far is conjectural. All assumptions are based on rationalizations only, since no data about the latter is available for comparison. Still, this study sheds good light on the subject and provides us with a better analytical toolkit to understand the market. CWPI, IPCA Labor, IPCA Repairs and IPCA Repairs w/o Labor (Index 6 = 100) 230 CWPI SP IPCA Labor IPCA Repairs IPCA Repairs w/o Labor CWPI, INCC, INCC Larbor, INCC Materials (Index 6 = 100) 230 CWPI SP INCC INCC Labor INCC Materials For more information, contact:: Gustavo Garcia Research Manager South America T +55 11 3513 6783 gustavo.garcia@sa.cushwake.com Jim Morrissey Americas Regional Research Director, MD T +1 (212) 713 6812 jim.morrissey@cushwake.com Cushman & Wakefield (C&W) is known the world-over as an industry knowledge leader. Through the delivery of timely, accurate, highquality research reports on the leading trends, markets around the world and business issues of the day, we aim to assist our clients in making property decisions that meet their objectives and enhance their competitive position. In addition to producing regular reports such as global rankings and local quarterly updates available on a regular basis, C&W also provides customized studies to meet specific information needs of owners, occupiers and investors. C&W is the world s largest privately-held commercial real estate services firm. Founded in 1917, it has 230 offices in 60 countries and more than 13,000 employees. The firm represents a diverse customer base ranging from small businesses to Fortune 500 companies. It offers a complete range of services within five primary disciplines: Transaction Services, including tenant and landlord representation in office, industrial and retail real estate; Capital Markets, including property sales, investment management, investment banking, debt and equity financing; Client Solutions, including integrated real estate strategies for large corporations and property owners, Consulting Services, including business and real estate consulting; and Valuation & Advisory, including appraisals, highest and best use analysis, dispute resolution and litigation support, along with specialized expertise in various industry sectors. A recognized leader in global real estate research, the firm publishes a broad array of proprietary reports available on its online Knowledge Center at: This report has been prepared solely for information purposes. It does not purport to be a complete description of the markets or developments contained in this material. The information on which this report is based has been obtained from sources we believe to be reliable, but we have not independently verified such information and we do not guarantee that the information is accurate or complete. Published by Corporate Communications. 2015 Cushman & Wakefield, Inc. All rights reserved. Cushman & Wakefield South America Praça Prof. José Lannes, 40 4º andar. São Paulo, SP. Cep: 04571-100 - Brazil 6