Real estate investment decision making – a review (2024)

Real estate investment decision making – a review

Article Type: Practice briefing From:Journal of Property Investment & Finance, Volume 30, Issue 5

Introduction

Over recent decades the asset class real estate has become increasinglyimportant throughout the world. This trend is shown through the increasingprofessionalism in asset management. For third parties, it can be very difficultto understand why an investor selected a specific investment over another. Theinvestment criteria in terms of risk and return appear to be obvious, but stillthe whole decision process is relatively opaque for an outsider.

The paper will analyse what decision models are available for real estateportfolio managers. It will introduce the reader to basic concepts of portfoliomanagement and their applications to real estate. The paper will examine if thetheoretical decision models (normative) and the practical real life decisions inportfolio management match in the UK property market.

General decision making theory

Generally speaking decision theory is concerned with the analysis ofjudgements. This theory seeks to lead to a profound and rational decision andderives from a variety of subjects, such as philosophy, mathematics, economicsand social sciences, including psychology and sociology (Ricciardi and Simon,2000). Generally speaking decisions are very important for every responsibleperson throughout life. In particular, the business and investment world ishighly dependent on responsible decision making. In order to come to a rationaland profound decision a decision analysis should be undertaken.

Howard (1988) describes decision analysis as “a schematic procedure fortransforming opaque decision problems into transparent decision problems by asequence of transparent steps”. Hence, this kind of analysis should assistthe applicant to overcome confusion and to get a clearer understanding ofpossible outcomes. Although the basic idea of decision analysis was establishedmany centuries ago most individuals often do not apply a proper decisionanalysis. The main reason for that lies in human nature. People generally areinured to make use of heuristics and intuitive reasoning (French and French,1997).

One way to overcome this problem is the establishment of models which eitherdescribe how decisions are made or give specific instructions on how to come toa decision. A further decision model uses both methods. Therefore, decisionmodels can be distinguished into three categories:

1. 1.

   Normative models.

2. 2.

   Descriptive models.

3. 3.

   Prescriptive models.

Normative models focus on the question how decisions should be made. Hence,they give investors a framework as to how they should come to a rationaldecision. In order to simplify the model it is assumed that people actcompletely rationally in accordance to their utility function.

By way of contrast, descriptive models analyse how in practice thesedecisions are made (French, 2001). This is possible through the application ofempirical research.

If the new findings of empirical research are incorporated in a model it islabelled as a prescriptive model. This model is there to overcome the gapbetween the normative and descriptive models. The main attribute of this modelis that it allows findings of descriptive studies to be taken into account in anew normative model. This is possible because the decision maker can adjustfindings of previous descriptive studies in his/her sensitivity analysis.

Property decision making theory

In order to illustrate the most important actions during a typical investmentdecision process the following steps should be highlighted (according toHartigay and Yu (1993)):

1. 1.

   Definition of purpose and objectives of the real estate portfolio.

2. 2.

   Formulation of a specified strategy and their selection/assessment criteria.

3. 3.

   Assessment of individual projects against the predefined criteria andstrategy.

4. 4.

   Closer examination of projects that meet return and risk profile.

5. 5.

   Closing the investment decision and implementation of actions.

6. 6.

   Post auditing and optimising (link to step 1).

It should be noted that there are differences between the decision makingmodels throughout the literature. There are models which were specially createdfor a typical investment decision process and models which were designed for abroader application. The four stage models of Roulac (1994) and Brown andMatysiak (2000) derive from capital budgeting rather than from investmentdecision making. The difference between both budgeting models is that Roulacprovides a decision phase, whereas Brown and Matysiak offer an implementation(3) and auditing phase (4) but no actual decision phase. In this context, it isto be noted that several models are missing important decision making steps. Ina similar way to Roulac (1994) the models of Jaffe and Sirmanns (2001) and Baum(2002) both stop the decision process after an investment decision has beenmade.

All of the above models have a sequential structure. Only a few authorsrecommend that their decision process should have a circular structure. Hartigayand Yu (1993) indicated the need for a circular structure, which is suggested intheir last step (conclusion, revise goals). Figure 1 should show an idealdecision making process which should be comprehensive but still connected to themost important grouping activities.

The so-called “definition phase” should define the investmentstrategy of an investment fund. Hence, this phase should establish tangibleobjectives, mostly in regard to the risk and return. Farragher and Savage (2008)notice that, from an empiric point of view, many investors tend to have notquantified their risk standard. However, it must be noted that this phase isessential for the further investment process so it is crucial that all detailsare precisely outlined and communicated throughout the organization. Thefollowing phase is the “planning phase” which generally tries todefine the previously established objectives more precisely. Therefore, theconcrete guidelines regarding the analysing and searching phase are established.Further on this phase includes an evaluation of the general environment,including a market overview. By way of contrast, the “dealing phase”is concerned with the search for and evaluation of suitable investment projects.It includes the forecasting and evaluation of all relevant investment criteria(return, risk, legal, tax, etc).

The executing phase is mainly focused on the actual decision making. Thisphase contains the final approval of an investment board. Only if all relevantcriteria are fulfilled a project will pass this phase. If this is not the casethen the process will be regressive in order to adjust the predefined andplanned objectives or to start the investment search again. This phase alsoinvolves the deal structuring and negotiation.

The next step is the watching phase which is mainly concerned with themeasurement and controlling of the investment. This phase then leads to theoptimising phase which will analyse the strengths and weaknesses of theinvestment property. This phase may either lead to an advancement of theinvestment property or, conversely, to a disposal. It is important to highlightthat this phase is interrelated to the first phase which emphasizes thecircularity of the investment process. Whether the investment was successful ornot, the decision will be evaluated to ascertain if the definition phase needsany adjustments for future.

Definition of portfolio management

Portfolio management is concerned with the asset allocation of the overallportfolio. Here, the fund manager can choose between various types of assetssuch as stocks, bonds, real estate and other investments. Formulating objectivesand an asset allocation policy with regard to the portfolio is a key task of theportfolio management (Nieboer, 2004). With regard to the real estate allocationthe portfolio manager can choose between various property types (commercial,offices, industrial, etc).

In contrast asset management focuses more on the management of the individualassets, meaning it focuses on decisions relating to a property such as therefurbishment, improvements and disposals. This operation also focuses onperformance analysis, pricing and marketing of properties. Generally speakingthe objective of real estate asset management (REAM) is the successfulapplication of the predefined strategy of the portfolio level (Fuerst, 2009).

Dubben and Sayce (1991) state that property management deals with the day today management of properties. The property manager acts on the behalf of thelandlord and has certain tasks to fulfil which can include rent collection,lease management and arranging maintenance and repair works. To distinguish thisactivity from asset management, the property manager generally does not considerthe performance (return) of the property.

Objectives of portfolio management

One of the first things to do in portfolio management is to set outobjectives. These can be either in accordance with company regulations(liquidity requirements, etc.) or with the preferences of the investor. Thesepreferences focus on variables like return, risk and liquidity. Time is a veryrelevant factor in regard to all three variables and the relation between theseobjectives must be analysed. Generally the variables risk and return follow thesame trend, by taking a higher risk the investor expects a higher return.Conversely, liquidity and risk follow contrary paths since the higher liquidityof an asset can contribute towards a lower risk. The time frame of an investmentmust also be considered. On a long-term basis the investor may be looking for asafe (low return and low volatility) investment. However, the investor may alsoopt for a riskier investment over a shorter time period.

Background of real estate portfolio management

Since portfolio management has a variety of different approaches, somedefinitions should be established. Stier (2010) concludes that it should bedistinguished between portfolio construction, portfolio planning and portfoliopolicy.

In context to portfolio construction the bottom up approach and the top downapproach must be mentioned. The bottom up approach is adopted to analyse anexisting real estate portfolio. Generally the approach examines how theportfolio can be positively restructured. Fuerst (2009) mentions that it is veryunrealistic to restructure historically grown portfolios according to abenchmark. The main reasons behind that statement are the characteristics of theasset real estate, mainly due to the lot size, illiquidity and transactioncosts, making it very difficult to match a benchmark.

On the other hand the top down approach is only applied when constructing anew portfolio on the basis of stochastic calculations. Since the portfolio iscreated from scratch it is generally possible to track benchmarks or to apply“efficient” allocation choices. Since most investors already have anexisting real estate portfolio the bottom up approach is more relevant inpractice.

Whilst talking about portfolio policy, one must be aware of the differencebetween the active and the passive approach. Wellner (2003) highlights that thepassive approach is based on the idea that the market is efficient. This meansthat an individual investor generally will not be able to beat the market in thelong run. This implies a buy and hold strategy which can generate appreciationsthrough inflation and positive market trends. The active approach is based uponthe opposite view which is that a good active management policy must be able tobeat the market.

Modern portfolio theory and other normative theories

Markowitz portfolio theory

Before Harry M. Markowitz introduced his mathematical approach in 1952 it wascommonly known that investors can benefit from naive diversification (Markowitz,1952). Hoesli and McGregor (2000) notice that main feature of the Markowitzportfolio theory (MPT) theory is that investors can increase their expectedutility (risk and return) by applying statistical diversification. Throughhaving the right proportion of each asset in a portfolio it is possible toexclude all suboptimal risk and return combinations. This process of increasingthe utility level is possible in the following two ways:

1. 1.

   The output can maximize the return in respect to a given level of risk.

2. 2.

   The output can minimize the risk by a given level of return.

In order to get such proportions of optimal asset weightings, the first thingto ascertain is to compute the expected risk and return of each asset. Furtheron it is to examine how each of the assets correlates to each other. Finally,through the application of Langrange multipliers it is possible to solve theoptimization problem (Amu and Millegard, 2009). The result can be highlighted asthe efficient frontier, which includes all optimal sets of portfoliocombinations.

One major criticism of the modern portfolio theory is that it “only”uses historical or estimated input data to examine the optimal allocation. It isobvious that fund managers need to find optimal asset combinations for thefuture and not for the past. On this subject, Sharpe (1990) noticed that the MPTis most useful for standard deviation and correlation analysis but relativelyuseless for expectation of returns (Hoesli and McGregor, 2000). Further on sinceasset allocation decisions are not made as often by investors it is worthmentioning that the theoretical “optimal” asset allocation can changeover time quite quickly.

Lee and Stevenson (2005) identify that possible results of optimalallocations might have extreme allocation results which are for a portfoliomanager difficult to follow. Also since these extreme allocations result frompast data, the requested diversification effect can be diminished, since timecan change the optimal allocation easily.

Hoesli and McGregor (2000) point out that the determination of investor’spreference can be difficult, since not all investors prefer to have the optimalrisk and return allocation and might also orientate on other issues likeliability matching. Discussing the theoretical framework it is obvious that thismodel is restricted to various assumptions which are not applicable in reallife. For example, the framework does not allow hedging vehicles like lending,borrowing and short sales which are generally good practices in portfoliomanagement.

The following issues are particularly relevant to real estate investments butsome of them might be also valid in general. First, the theory does notincorporate any transaction costs. In regard to stocks these costs might benegligible but the transaction costs in real estate are considerably higher. Inthis context, another problem is the illiquidity of real estate assets. The MPTtheory assumes liquid assets, since this model was first developed for the stockmarket. The real estate market is a relatively illiquid market and it takes timeand money to sell real estate assets.

Capital asset pricing model

The capital asset pricing model (CAPM) model is a supplement to the portfoliotheory, which was developed by Sharpe, Treynor, Litner and Mossin (Perold,2004). This theory examines the rate of return of an individual asset in regardto the market risk Beta. The theory introduced the concept of the securitymarket line (SML), which should give evidence how the market would price asingle asset in relation to the return and the systematic risk. Hence, thismodel verifies the important input parameter return for the MPT theory.

With regard to the CAPM it must be said that this method can be used for theestimation of discount rates. Here, one must point out that since not all realestate investors are public traded companies it will not always be possible tofind an appropriate beta figure. Baum (2001) notices that the beta factorgenerally implies the risk of the past for the estimation of future returns.However, since this method is widely used in the corporate finance world it isgenerally applicable for Real Estate Investment Trusts (REITs). In the generalreal estate practice they mostly rely on the estimation of discount rates fromempirical surveys.

Another important issue on this topic is the general availability of dataabout time and quality. Some developed markets, the US and the UK in particular,offer a good basis (according to JLL transparency index). In respect toindividual real estate assets it is very difficult to forecast appropriateinputs. General benchmark indices like the IPD all property index are appraisalbased, so therefore this data cannot be evaluated as market evidence.

All these factors indicate that the general portfolio theory can possiblygive good advice on asset allocation for a mixed asset portfolio. It is possibleto estimate what proportion of the investment budget should be allocated to realestate, so that diversification benefits can be applied. These theories are notdevised for real estate investments and they offer little benefit for assetallocations on the real estate level.

However, the real estate practice has several practical techniques to analysewhether an investment is preferable or not. The following part will give a briefoverview of the most relevant practical decision techniques.

Practical decision models

The following section will give, in contrast to the previous theoreticalmodels, an overview of the most relevant practical normative decision models.

Return on investment

The return on investment (ROI) is a static metric which is commonly usedbecause of its simplicity and comparability. As equation (1) illustrates, itdivides the net benefit of an investment with the costs of an investment. Asillustrated, the ratio is also defined as Earnings before Interest and Taxesdivided by total assets:

This method can be used for the fast and approximate evaluation of aninvestment. Nonetheless, this method does not take compound interest and thetime value of money into account.

Discounted cash flow approach

One of the most important factors in a property is the underlying cash flow.Generally speaking the discounted cash flow approach (DCF) model can value thecash flow over time and can estimate what value the property has currently. AsGeltner et al. (2007) notice the appraisal is based on the assumptionthat the market value is equal to the net present value (NPV) of the futurereturns. The following formula exhibits the NPV formula:

Whilst using this formula it must be highlighted that the exit value of thepotential sale of the property must be incorporated into the cash flow. In orderto estimate the NPV various data must be incorporated accurately into the model,particularly the discount rate (required rate of return), which is crucial forthe estimation of the NPV. It must be pointed out that many input parameters(rent, discount rate, exit yield, etc.) have a strong influence on the NPV, sotherefore, the proper estimation of these figures is crucial in order to obtaina reliable result.

Generally speaking it can be concluded that if the DCF has a positive NPV itcan be evaluated as a lucrative investment. The different investment propertiesshould be compared and the one with the highest NPV should be preferred (only inregard to this criterion). This rule is only applicable if the capital outlay ofeach of property is the same. The investor also has the option of adjusting theDCF parameters in order to make the results comparable. Nonetheless, it ispossible to compare the results of different capital structures by using thebenefit/cost ratio. This approach can be achieved “by dividing thediscounted present value of the total benefits to be obtained from a project bythe discounted present value of its total costs” (Enever et al.,2010).

However, by using the net cash flow it is possible to estimate the internalrate of return (IRR). The following section will highlight the advantages of theIRR as a return figure.

Internal rate of return

In comparison to the previous mentioned return figure (ROI), the IRR has theability to take the time value of money into account. It is defined “asthe rate of return which equates the present value of the cash outflows to thepresent value of the cash inflows” (Brown and Matysiak, 2000). Therefore,the NPV equation is equalled to zero in order to solve the rate of return (r) byusing an iterative technique:

c_t = cash out in t, a_t = cash in flow.

Although the IRR has some conceptual drawbacks such as the liability to thenumber of sign changes, it is still the most frequently used return figure. Onereason for that is if it is assumed that there is only one sign change, thismethod provides the opportunity to compare projects easily. The general decisionrule of this technique is that the project with the highest IRR should beaccepted if it exceeds the opportunity costs of capital.

The previous mentioned normative models can help investors to generatecomprehensive results regarding their investment selection. From a theoreticalpoint of view the mean variance framework can be helpful for the portfoliostructuring of a mixed asset portfolio. On the other hand the illustratedpractical techniques (DCF, IRR, etc.) are widely used tools for the evaluationof single projects in the real estate sector. Further advanced quantitativetechniques such as resampling and downside risks optimization are applicable forthe real estate sector and they are increasingly applied. Still it isquestionable if these models can explain the whole investment decision. A surveyof UK decision makers was undertaken.

The survey

First of all the general research question can be summarised the followingway:

What are the most relevant decision criteria/techniques for real estateinvestors in the UK and did their real estate investment decision approachchanged in recent years?

There are several reasons why this particular research question has beenformed. First of all it is very interesting to further examine what factors havea strong influence on a real estate investment decision. French (2001) alsoindicated that there is a need to further analyse the relatively opaque processof real estate decision making. As already outlined, many studies suggest astrong influence of intuitive decision behaviour, so that it would be worthwhileto analyse if this is still the case. Later studies such as Roberts andHenneberry (2007) suggest that the investment decision practice has beenmodified only rarely in the last decades. In respect of Farragher and Savage’swork (2008) it should be evaluated if the search phase (or a similar step) ofthe decision making process is still considered as most important. Their studyindicated that the phase of the property search (dealing phase) was consideredas most important by investors. Other studies such as Worzala (1997) andGallimore et al. (2000) indicated that investors rarely apply advancedquantitative models. Hence, it is questionable if this fact has changed throughthe financial crisis.

The main reason why the UK market was chosen lies in the reason that the UKis one of the most transparent and mature real estate markets, as the Jones LangLaSalle Global Real Estate Transparency Index 2010 displays[1]. Anotheradvantage of this market is that it is familiar and easily accessible for theauthor.

None of the previously analysed studies questioned a future trend of theasset allocation framework of investors. Most of the survey literature focusedon market sentiment (French, Gallimore and Gray), the decision process (Robertsand Henneberry, Farragher and Savage) and general decision criteria of realestate investments. The survey seeks to be a relevant supplement of thesestudies that will offer additional insights to the latest developments ofinvestment decision making.

The most important research questions are as follows:One important innovationof the survey is that it questions the relevant decision model at the stage ofthe dealing phase (analysing phase). Hence, the results should give answers asto which practical considerations are most important during this decision phase.

As mentioned earlier the survey received through the method of telephoneinterviews a response of 16 interviews from an initial sample of 78 investors,which represents a response rate of 21.8 per cent. The literature suggests thatthis result is a good average (Gallimore et al., 2000).

The survey is divided into the following three issues: general questions,decision making process and decision making in the economic cycles and thefuture.

General questions

Most of the respondents were “fund managers” of private investmentcompanies. The average work experience of the respondents was between five andten years. The average size of a portfolio (mixed assets and real estate (RE))was above £3 billion, with an average real estate portfolio of £2.933million. The strongest proportion of the asset allocation regarding propertytype was the retail sector (with 36 per cent).

Decision making process and investment decision criteria

The survey asked about the importance of each decision stage. Hence, atypical investment decision process was presented to the respondents.

On the subject of the general ranking, some respondents indicated thefollowing: “It is very hard to rank these stages, since they areinterlinked they can’t be pursued without the examination of the previousstage”.

Nonetheless, the respondents on average ranked the “dealing phase”as most important. They indicated that the search and analysis of potentialinvestment properties is, from their point of view, the most important stage. Incontrast the “planning phase” was ranked as least important. Most ofrespondents mentioned that they generally do not take too much time for theplanning phase, since this phase should just further outline the definitionphase.

With regard to the definition phase all respondents answered that they havecertain objectives to define which are generally related to the risk/return andother standards.

The next question asked the respondents about the most important decisiontechnique and criteria in the dealing phase. The respondents had the option tochoose between more quantitative or qualitative models and were able to indicateone from each type. Nearly, all respondents chose to pick both criteria (14 outof 16) with only two candidates picking from only one criterion. As Figure 2indicates most of the respondents used practical techniques (DCF, IRR, etc.) asa quantitative model (81.25 per cent). On the other hand the most importantqualitative criteria were the property specifics (62.5 per cent) and the generalexperience (25 per cent). Important to highlight is that only one InvestmentAnalyst selected the MPT. Several fund managers indicated that they know ofthese theories but they do not make any use of either MPT or CAPM.

Future trends and the economic climate

The initial questions covered the issue of decision making with reference tothe economic climate and future trends. The later questions asked therespondents about their expectations as to which model will become moreimportant in the near future (five years). Many fund managers outlined thatsensitivity analysis and risk models will receive more importance. Most of themnoted that the both are combined. Many participants mentioned that they aregenerally happy with their risk assessment tools, but they are still seeking toexpand and improve these models.

The next question asked the respondents if the proportion of real estate theymanage has dramatically changed through the onset of the economic crisis. Atthis point it is worth mentioning that the question only makes sense forrespondents who have a mixed asset portfolio. However, by just looking at therespondents who meet this criteria, most of them indicated that the proportiondid not dramatically change. This question was aiming to identify if real estatecould be seen as a relatively secure investment compared to equity stocks. Sincethe result is not strongly angled in either direction, it can be concluded thatmost of the portfolios did not experience a dramatic decrease in the real estateproportion.

In this context, many participants outlined thoughts similar to this:

Certainly the value of the real estate assets and the whole fund narrowed,nonetheless through cash holdings and other measurements we were generally ableto hold our real estate assets.

Another question asked the sample group if there was any decision model(quantitative of qualitative), which played the major role during the economicboom. Most of the investors (12 out of 16) indicated that they made more use ofqualitative models, especially their general experience, during this time. Onthe other hand some respondents (three people) highlighted that they used acombination of quantitative and qualitative models. One life fund manager saidthe following:

As mentioned both models are relevant, however nowadays greater risk analysisand stress test might be more important.

In contrast two fund managers emphasised the importance of practicaltechniques (DCF, etc.) and the general experience (qualitative). They indicatedthat through troubled times, general experience will be a key characteristic togenerate stable returns.

Finally, the participants were asked if their views have changed through theappearance of the economic crisis. The result shows that most of the respondentshave changed their view (11 people vs five). This result in combination with theprior questions emphasises the idea that risk assessment tools and otherquantitative techniques have become more relevant. Before conducting the surveyit was questionable whether real estate investors have changed their decisionmodels from a more intuitive way (qualitative models) to more quantifiablemodels (quantitative models). However, this implication could be partiallyconfirmed which does not however mean that the general experience does notcontinue to have a high importance in the investment process. Nonetheless, theresult indicates that investors are still highly reliant on their generalexperience but make stronger use of advanced risk assessment methods which, whentaken in combination, could be seen as an advancement.

Summary

This survey was seeking to garner some recent insights into the decisionmaking practice of high value UK real estate investors. The sample included allmajor London based investment companies with an average real estate value ofabove £3 billion. The respondents were all highly qualified with an averageprofessional experience of between five and ten years. Generally all the realestate portfolios had a focus on retail properties.

Regarding the decision making process the survey highlights that the dealingphase (analysing phase) is the most important decision phase. On this subjectthe respondents highlighted that both the practical techniques and propertyspecifics play a major role. In the overall decision process the practicaltechniques, the general experience and benchmarking are seen as the mostrelevant for an investment decision. It is a matter of some note that only oneinvestment analyst highlighted the relevance of the MPT theory as a decisionmodel. Nonetheless, most of the respondents indicated that sensitivity analysiswill be one of the most important tools for the future. Also most therespondents indicated they are going to expand their risk assessment. During theeconomic boom most investors favoured qualitative criteria, in particulargeneral experience, as the most important decision model. Nowadays this view haschanged since most of the fund managers indicated a stronger focus on riskmodels and quantifiable values which does not necessarily have to imply theexclusion of the general experience.

This research gives further insights into the investment decision makingprocess of real estate investors. The work explained the DM process in generaland in context to real estate. Further on the thesis examined various normativemodels that can support the investment decision. Theories like the MPT give agood background for professional portfolio management, but techniques such asDCF, IRR, down side risk minimization and resampling are much more applicablefor the real estate sector.

The empirical survey gives the latest insights into the development of the DMcontext of UK investors. Hence, it must be stressed that the general experienceand practical techniques are still most important for the DM process. Investorsalso evaluated the analysing phase (dealing phase) as the most importantdecision phase. Although the MPT is rarely applied in practice, most investorshighlighted their advancement in quantitative techniques, especially thesensitivity analysis and downside risk models. In contrast to the givenassumption of the literature the practice has advanced in some way. Obviouslymost investors revaluated their decision models and nowadays they have stricterand more tangible standards (in context to risks, etc).

Henry Bispinck

Note

1. UK achieved the third rank in the JLL Global Real Estate TransparencyIndex 2010.

Abbreviations: RQ1. What investment decision modelsare most relevant for real estate decision making?; RQ2. Has the use ofdecision models changed in the recent years?; RQ3. Have the applieddecision models changed either during or after the financial crisis?; RQ4.During a typical real estate investment decision process, which stage can beseen as most relevant?; RQ5. Which decision models will become moreimportant in the near future?

Acknowledgements

The author would like to especially thank his supervisors Claire Roberts andNick French. Without their advice and their professional network this work wouldnot have such an empirical value. The author would also like to thank hisparents for their tremendous support during his studies.