Demand forecasting in Real estate sector for housing : Case Study : DLF housing projection for Gurgaon

	
	

Dec 12,2012

Real estate as the name suggest is considered the real asset class world over and has been considered an asset class which always stood the value of time and never lets its investors down and this adage holds much stronger belief in India than anywhere else. However one of the worst economic slowdowns in the world have been driven by boom in real estate sector be in East Asian crisis of late 90s, Japanese decade of low growth after the real estate boom of early 80s or recent global turmoil led by American housing industry fall. India has not been an exception as when analysts and experts are trying to beat each other in betting on Indian real estate, 2008-2009 suddenly put brakes to those estimates and suddenly Unitech and DLF the Indian real estate bell weathers looked like out grown kids with jelly filled knees and for a change some people realised that real estate sector too can face corrections and suddenly people started talking of various factors which drive demand of real estate sector in a micro market and its not GDP alone. In this background an attempt has been made to identify various factors which affect real estate demand for a micro market. Given our familiarity with Gurgaon and it being the front face of IT/ITES sector which is considered the face of India represent most vibrant market and does not suffer from constraints like lack of supply or distance from a major market. Gurgaon has it all, poster boy of IT/ITES sector, close proximity to national capital, upcoming infrastructure and catchment of a large educated workforce. DLF has been chosn as its is India’s largest real estate company and synonymous with building of Gurgaon and Gurgaon is not about an attempt by a govt body to create a new city but just a function of vision and courage shown by DLF promoters more than 30 years back.
This study aims to capture the demand drivers in Gurgaon micro market and its assumed that since DLF is largest player in this market it will have a linear relationship with the demand seen in the Gurgaon market so an attempt has been made to estimate demand on the basis of various factors and historical patterns and see its relationship with deman

1. Objective & Relevance

The aim of the study is to estimate demand for DLF for the year 2011 after analyzing various factors and by using various economic tools by developing a model. The study aims to use regression analysis in order to understand the relationship and identify the pattern.

This study is highly relevant as it aims to estimate demand of housing industry using economics models and give an insight to a working executive about real patterns impacting demand rather than just simple model of build whatever one can do – a model which has gone horribly wrong in US and has almost build oversupply in various Indian cities as well. Further real estate sector though much more organized than before is still marked by plethora of small builders who aim to build depending upon their financial strength rather than using any analysis to estimate demand and then work according to rule of demand and supply and hence this study is going to be very beneficial to these industry professionals.

 

Embedded Scribd iPaper - Requires Javascript and Flash Player
Project Report : Real Estate Forecasting Demand Forecasting of DLF for FY
ACKNOWLEDGEMENT
We would sincerely like to thank our mentor and teacher who instilled in all of us the curiosity and interest for Economics and constantly ensured that we are on our toes as far as the lessons and subject matter grasp is concerned. The quizzes, the presentation and now this report though harassed, tormented, agonized and surprised us but we knew the real purpose behind it; building and nurturing our understanding of managerial economics concepts. Thank you Sir for being supportive, accommodating, encouraging, taxing and demanding. This project would not have been completed without your support. We are lucky to have been under your tutelage.
TABLE OF CONTENT
1. Introduction 2. Objective & Relevance 3. Review of Literature 4. Demand Forecasting Methodology
Objective of the study Review Literatire Support A clear methodology – process Analysis and finding Business decision Recommendation Conclusions Annexures 1. Brief profile of the firm 2. 2. Primary table for data collected with source 3. 1) Indian Real Estate industry 2) Introduction to DLF 3) 4) Demand Forecasting Model 5) Interpretation & Analysis 6) References
1. Introduction Real estate as the name suggest is considered the real asset class world over and has been considered an asset class which always stood the value of time and never lets its investors down and this adage holds much stronger belief in India than anywhere else. However one of the worst economic slowdowns in the world have been driven by boom in real estate sector be in East Asian crisis of late 90s, Japanese decade of now growth after the real estate boom of early 80s or recent global turmoil led by American housing industry fall. India has not been an exception as when analysts and experts are trying to beat each other in betting on Indian real estate, 2008-2009 suddenly put brakes to those estimates and suddenly Unitech and DLF the Indian real estate bell weathers looked like out grown kids with jelly filled knees and for a change some people realised that real estate sector too can face corrections and suddenly people started talking of various factors which drive demand of real estate sector in a micro market and its not GDP alone. In this background an attempt has been made to identify various factors which affect real estate demand for a micro market. Given our familiarity with Gurgaon and it being the front face of IT/ITES sector which is considered the face of India represent most vibrant market and does not suffer from constraints like lack of supply or distance from a major market. Gurgaon has it all, poster boy of IT/ITES sector, close proximity to national capital, upcoming infrastructure and catchment of a large educated workforce. DLF has been chosn as its is India’s largest real estate company and synonymous with building of Gurgaon and Gurgaon is not about an attempt by a govt body to create a new city but just a function of vision and courage shown by DLF promoters more than 30 years back. This study aims to capture the demand drivers in Gurgaon micro market and its assumed that since DLF is largest player in this market it will have a linear relationship with the demand seen in the Gurgaon market so an attempt has been made to estimate demand on the basis of various factors and historical patterns and see its relationship with deman 2. Objective & Relevance The aim of the study is to estimate demand for DLF for the year 2011 after analyzing various factors and by using various economic tools by developing a model. The study aims to use regression analysis in order to understand the relationship and identify the pattern. This study is highly relevant as it aims to estimate demand of housing industry using economics models and give an insight to a working executive about real patterns impacting demand rather than just simple model of build whatever one can do – a model which has gone horribly wrong in US and has almost build oversupply in various Indian cities as well. Further real estate sector though much more organized than before is still marked by plethora of small builders who aim to build depending upon their financial strength rather than using any analysis to estimate demand and then work according to rule of demand and supply and hence this study is going to be very beneficial to these industry professionals. 3. Review of literature
Theoretical Basis of Estimation The methodology adopted in theoretical basis of estimation follows the approach formally rendered by Rosen (1974), setting out a model of demand, supply and competitive market equilibrium. A housing unit is described by a vector of n objectively measurable characteristics. The housing unit (a bundle of characteristics) commands a price in the market. Various bundles and their associated prices reveal the implicit prices of characteristics, known as hedonic prices. Empirically, these prices can be determined from regressions of price on characteristics. This model has been used to generate the housing price indices and to estimate demand for housing characteristics (see the survey in Quigley, 1979). The demand for housing depends on the characteristics of the household, its income and the price of the house. The income that decides the purchase of a house is not only present income but also the expected future income. The prices that are taken for demand analysis are the housing price indices that are estimated using hedonic price analysis for the housing characteristics of various regions. The demand is measured in terms of household expenditure on housing. The methodologies to estimate the demand for housing characteristics and the household's housing demand are discussed below. 3.1 Estimation of Demand for Housing Characteristics Let Z = (Zi.... Zn) be a vector of housing characteristics and P(Z) be a hedonic price function defined by some market clearing conditions. The household decision is characterised by the utility function U = U(X, Z), where, X is a composite commodity other than housing whose price is unity. Households maximise utility subject to their budget constraint Y = P(Zi) + X. The first-order condition yields, Pi = U(Zi)/U(X), where, Pi is the implicit price that households have assigned to characteristic i. Estimation of these implicit prices can be done by regressing market values of house prices P, measured as rents, as a function of various housing attributes. Thus P(Z) = f(Z, ...,Zn).
Most of the analyses in developing countries suggest that a semi-log specification is a good approximation to the best-fit functional form (Follain and Jimenez, 1983). The marginal willingness to pay for a particular characteristic can be interpreted as the derivative of the hedonic regression with respect to that characteristic. 3.2 Estimation of Housing Demand Conventional demand analysis postulates a relationship between the quantity of good demanded, its relative price, the income of the household and other household characteristics affecting demand. In this perspective, one can identify Qh = f(Ph, Y, H1, H2, ...) where, Qh is the quantity of housing services demanded, P h is the relative price of housing; and Hi (i = 1 ... I) are the household characteristics. With regard to housing demand analysis, there are crucial issues of measurement of the variables, particularly Qh, Ph and Y, as well as functional form of the equation. In the literature on the measurement of the dependent variable related to the quantity of housing consumed, it is common to use expenditure on housing, in the form of rent paid, as a surrogate for consumption (Malpezzi and Mayo, 1985). The rent R is, however, a product of unit price and quantity consumed; and the relation postulated above becomes R = Ph Qh = f(Ph, Y, Hi). Depending upon one's notion of the housing services bundle, the price variable will vary for each household if it is related only to shelter size, or it will be constant for a sub-market, when housing is defined broadly to include shelter, neighbourhood characteristics and access. Consistent with the hedonic equations approach, the price variable has been included as constant for a sub-market. The problems associated with measurement of housing consumption and house price are closely related. In an earlier equation, it was suggested that the housing price variable, Ph, can be constant for a sub-market. Most studies of housing demand have assumed a constant price throughout the sample and hence have omitted the price term. This omission may bias other demand equation parameters if the omitted price term is correlated with other variables. This is particularly true if a single linear expenditure equation is used for estimation. A few research studies that include the price variable (for example, King and Mieszkowski, 1973, and Polinsky and Elwood, 1979) have allowed the intra-metropolitan prices to vary across neighbourhoods or by dwelling units. Such a specification may again wrongly state the price variations in the sample, if households were limited in their choice to specific neighbourhoods or dwelling units. Mayo (1983) in a study for Egypt, uses land price variations as a surrogate for housing price. In this study, Ingram's (1987) perception of identifying a housing price variable on the
basis of available information regarding the various market segments is followed. The hedonic analysis, presented earlier, was undertaken for these sub-markets. Multiple line equation(s) cannot be represented in ASCII text where, Pij is the market price of ith tenure class in the jth zone; and Xk are the dwelling characteristics. A sample-wide average of the housing attributes as representing a standardised bundle of housing services is calculated and the price of this bundle for each zone is estimated. The next important issue is of the functional form of the demand equation. Linear expenditure functions are most often used in modelling based on demand equations, where the functional form satisfies certain criteria required for the system-wide consistency of price and income elasticity estimates. The linear function is, moreover, consistent with the StoneGeary utility function-implying that it has theoretically appealing properties and permits explicit derivation of the parameters of the underlying utility function (Mayo, 1978). The linear form is, however, restrictive in the sense that income and price elasticities measured through it are constrained to increase monotonically as prices and incomes change. The loglinear form, on the other hand, provides estimates of constant elasticities that are independent of levels of income, prices or demographic variables. 3.3. Variable considered for Regression Variables Considered for Estimation of Hedonic Function Estimations The idea behind hedonic analysis of demand for housing characteristics is that the price of a dwelling unit is a function of various characteristics of the house such as the shelter quality, size, dwelling characteristics, etc. House prices. The distinction between the stock of housing and the flow of services it yields over a period of time provides two measures of value, commonly encountered in the housing market. These are rent, which is the payment made for a flow of housing services received over a specified period of time, and price, which is capital value associated with a particular unit of stock. Information regarding the capital value of stock is not available in the NSS household survey data. Rents reported by households are used in the regression analysis of tenants as well as owners. For owned houses, the rentals are imputed rents. Shelter-related variables. The area of a dwelling unit and the type of building (flat, chawl/bustee, independent house, etc.) are two of the important variables which determine the rent of the housing unit. The type of construction (pucca, semi-kuccha and kuccha), the type of flooring (pucca if constructed using bricks, cement and stone; kuccha if constructed using wood, bamboo and reed) and the housing condition (excellent or otherwise) determine the rent of the house.
Variables Considered for Housing Demand Function Estimation The demand for housing, measured in terms of housing services, of a household depends on the characteristics of the household and of the neighbourhood. Housing expenditure. The demand for housing is measured in terms of expenditure that households incur on housing services. The actual rent paid by tenants and imputed rent paid by the owner-households is a measure of housing expenditure. The higher the housing expenditure, the higher is the household consumption of housing services. Household characteristics. The age of the household head determines the differences in life-cycle pattern of housing consumption. There is a non-linear pattern of age and hence the age enters the demand equation linearly and quadratically (Borsch-Supan, 1987). Another demographic variable that affects housing demand is household size. Household size is one of the important variables determining household expenditure on housing. This variable is expected to have a positive sign indicating that as household size increases, households move to a bigger house. A useful measure of expenditure on housing is the crowding of a dwelling in terms of the deviation of the actual dwelling size from some optimum (Borsch-Supan, 1987; Behring and Goldrian, 1987; Booth, 1976). The present paper adopts the following measure of crowding: Crowding = (AROPT- ARACT)2 AROPT = 45 + (15 x MEM - 45)/2 where, ARACT is actual area; AROPT is optimal area; and MEM is household size. Income. Household income is the main criterion determining household demand for houses. The measure of income considered in the present analysis is the annual income of the household. Price. The price of the houses is difficult to measure because of the heterogeneous nature of the dwelling units. Hedonic price indices are used as a measure of price for a standardized bundle of housing for different areas. It is assumed that the price of this standardized bundle will remain the same for that state.
Demand Forecasting Methodology We estimated various factors which impact demand of luxury housing in any micro market and more so Gurgaon and chose following parameters : 1. Growth in IT/ITES salary : rather than taking salary in absoluter term , we focussed on growth as growth signifies more of a disposable income as a person expenditure
do not increase so much as per salary growth. Further we considered only IT/ITES sector as no of professionals in higher income bracket are much more in this sector compared to any other industry and hence linear relationship with the demand. 2. GDP Growth : GDP growth signifies overall activity in economy, confidence level and overall vibrancy in the system and since housing purchase is very strong well thought personal decision, emotional comfort and outlook over certainty of income plays a very important role and hence GDP was considered as an important tool. 3. Price of apartments : Like income level, ability to afford a house in a particular segment is very important and hence apartment cost was taken as also one of the factor impacting demand as there is elastic relationship between price and demand of apartments. 4. Home loan rate : Since 90% of sale in NCR region is backed by mortgage, change in interest rate makes the difference between affordability and non affordability and hence its factor which impacts housing sales drastically. 5. Absorption of Commercial space in Gurgaon: Gurgaon despite being a old town has not much of its own catchment area and consists of mainly villages. Though lot f auto companies are there in vicinity but they don’t drive market for luxury housing on account of less no of people in senior management and a different demography. Hence its the ITES/IT sector which is driving all the employment growth in the city as well as demand for housing as more an more people come here every year. Further it was assumed that a over a larger period of time, people prefer to buy apartments closer to their offices as well as centre of economic activity and hence housing demand will be driven by economic activity rather their actual living and hence we considered absorption of office space in city as important parameter.
We gathered that the primary factors driving housing demand were Growth in IT/ITES Industry, Rising Disposable income, Easy availability of Mortgage and Growing urbanization. We based our parameters the (x’s) therefore on growth in salary, multi-story apartment cost per sq ft, home loan interest rate, GDP growth & Absorption of Commercial / office Space in Gurgaon. The last factor was the result of Gurgaon being an increasingly attractive destination for Multi-national and Pan National organizations and the likelihood of continuous increase in demand here. Data Source : Growth in IT/ITES salary as been taken by NASSOM surveys and various IT salary surveys carried out in business magazines and website like SilioconIndia, Dataquest, Business Today , NASSCOM and Executive search firms.
Real estate cost data for Gurgaon from websites such as magic bricks, DLF, HUDA.
Home Loan Interest Rate from: Home loan rates have been taken on the basis of interest rates charged by leading financial institutions such as ICICI, HDFC, SBI. Demographical statistics from: IMRB report on Demographic profile and settlement pattern of NCR & www.euromonitor.com (India statistics) Absorption of Commercial / office Space in Gurgaon: India Office Market view CB Richard Ellis Report, DTZ research report. We found data available on these parameters from 2008 Q1 to 2010 Q1. This helped us in estimating the ‘y’ for each quarter i.e. Sale of space by DLF in Gurgaon. Finally, we performed Regression analysis of collected data and reached to our forecasting conclusion. Demand Forecasting Model
We conducted multiple regression using the formula: Y=m1X1+m2X2+m3X3+m4X4+m5X5+b Where y: Sale of Space by DLF in Gurgaon (Sq Ft) X1: Growth in Salary; X2: Multi-story Apartment cost per Sq Ft; X3: Home loan Interest rates; X4: GDP Growth; X5: Absorption of Commercial Space (Sq Ft).
Data
Sale of Space by DLF in Gurgaon (Sq Ft) 300,000 240,000 50,000 50,000 200,000 800,000 1,350,000 900,000 990,000
FY 2008 Q1 2008 Q2 2008 Q3 2008 Q4 2009 Q1 2009 Q2 2009 Q3 2009 Q4 2010 Q1
Growth in Salary 6% 6% 3% 1% 1% 1% 2% 4% 5%
Multistory Apartment cost per Sq Ft 5700 5400 5200 4800 5200 5300 5300 5200 5300
Home loan Interest rates 9.00% 9.50% 9.75% 10.00% 10.50% 10.00% 9.00% 9% 8.50%
GDP Growth 6.47% 6.47% 6.47% 6.47% 5.68% 5.68% 5.68% 5.68% 8.78%
Absorption of Commercial Space (Sq Ft) 800,000 500,000 250,000 100,000 700,000 900,000 500,000 500,000 700,000
On doing regression analysis ( liner regression with multiple variables on Ms Excel Sheet) we came with following results
m5 m4 m3 m2 m1 b 0.552040681 -860608.7649 -75295998.45 318.895662 -20126449.72 6352843.407 0.779573642 14881648.72 20626534.26 1234.616368 12386279.08 6503841.188 0.845776724 304122.5975 #N/A #N/A #N/A #N/A 3.290463343 3 #N/A #N/A #N/A #N/A 1.52168E+12 2.77472E+11 #N/A #N/A #N/A #N/A
Hence the value of equation Y=m1X1+m2X2+m3X3+m4X4+m5X5+b And result of Y is
Growth Cost of in apartment salary ( sq ft) 6% 6% 6% 6% 5500 5500 5800 5900 Commercial space absorption Demand for DLF 900,000 1,000,000 1,000,000 1,000,000 920,297.9 599,021.9 694,690.6 343,903.8
Year 2010 Q2 2010 Q3 2010 Q4 2011 Q1
Home Loan 8.50% 9.00% 9.00% 9.50%
GDP Growth 8.78% 8.78% 8.78% 9.50%
Regression analysis ( using MS Excel)
X X -(2,000,000) -1000000 1000000 2,000,000 0 SUMMARY OUTPUT 20.00% 10.00% 0.00% 6000 4000 - 6… 9… 5… Regression Statistics Multiple R 0.919661201 R Square 0.845776724 Adjusted R Square 0.588737932 Standard Error 304122.5975 Observations ANOVA df Regression Residual Total 5 3 8 SS 1.52E+12 2.77E+11 1.8E+12 Standard Error MS 3.04E+11 9.25E+10 F 3.290463 Significance F 0.177897974 9 R Y
X
Coefficients
t Stat
P-value
Lower 95%
Upper 95%
Intercept X Variable 1 X Variable 2 X Variable 3 X Variable 4 X Variable 5
6352843.407 -20126449.72 318.895662 -75295998.45 -860608.7649 0.552040681
6503841 12386279 1234.616 20626534 14881649 0.779574
0.976783 -1.6249 0.258295 -3.65044 -0.05783 0.708132
0.400714 0.202652 0.812886 0.035481 0.95752 0.529925
-14345281.95 -59545117.81 -3610.204637 -140938836.2 -48220656.73 -1.928910576
27050968.76 19292218.38 4247.995961 -9653160.708 46499439.2 3.032991938
RESIDUAL OUTPUT Standard Residuals -1.46198 1.640324 -0.5277 -0.54877 -0.22177 0.214351 1.390925 -0.71426 0.228875
PROBABILITY OUTPUT
Observation 1 2 3 4 5 6 7 8 9
Predicted Y 572272.9948 -65487.90046 148276.2922 152200.9234 241302.414 760080.1086 1090959.323 1033020.755 947375.0888
Residuals -272273 305487.9 -98276.3 -102201 -41302.4 39919.89 259040.7 -133021 42624.91
Percentile 5.555555556 16.66666667 27.77777778 38.88888889 50 61.11111111 72.22222222 83.33333333 94.44444444
Y 50000 50000 200000 240000 300000 800000 900000 990000 1350000
1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0% 1% 2% 0.06
300000 Predicted 300000
0.06 Line Fit Plot
3%
4%
5%
6%
7%
300000
(200,000)
1,600,000 1,400,000 1,200,000 1,000,000 800,000 300000 600,000 400,000 200,000 4700 (200,000) 4800 4900 5000 5700
300000 Predicted 300000
5700 Line Fit Plot
5100
5200
5300
5400
5500
1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0.00% (200,000) 2.00% 4.00% 0.09
300000 Predicted 300000
0.09 Line Fit Plot
300000
6.00%
8.00%
10.00%
12.00%
1,600,000
0.0647 Line Fit Plot
1,400,000 1,200,000 1,000,000 800,000 300000 600,000 400,000 200,000 0.00% (200,000) 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 8.00% 9.00% 10.00%
0.0647
300000 Predicted 300000
1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 (200,000) 200,000 400,000 800000
300000 Predicted 300000
800000 Line Fit Plot
300000
600,000
800,000
1,000,000
Interpretation & Analysis Demand for real estate of DLF in the period (2010-Q2-2011-Q1) is likely to decline in major parts barring a marginal increase in Q4 2010; 2010 Q2 has declined from the last quarter of 2010 Q1 from 990,000 to 920,297; It will further decline in 2010 Q3 to 599,021.9 from 920, 297.9 sq ft, and then decline from 21 Q3; A marginal recovery in demand is projected for Q4 2010 from 599,021.9 to 694,690.6 The demand in Q1 2011 will substantially decline as compared to the previous quarters of 2010; also it will the lowest demand period across quarters under study. Reasons: ‘Absorption of commercial space in Gurgaon is likely to rise in Q3 2010 and remain constant till 2011 Q1; Interest rates are expected to rise in Q3 2010 from 8.50% to 9% & further likely to increase to 9.50% in Q1 2011 after remaining constant in Q4 2010.
Annexure 1
Indian Real Estate Industry
This year has witnessed an above-average economic growth in India. Strong population growth, a large pool of highly-skilled workers, greater integration with the world economy and increasing domestic and foreign investment are expected to further drive India’s real GDP by 6% p.a. over the next 10 to 15 years. It is the services outsourcing which is revving up office demand. India is the prime destination for IT services outsourcing. In the coming five years, at least 55 million m² of extra office space must be completed in the premium office segment alone. By 2030 India will need up to 10 million new housing units per year. Rapid population growth, rising incomes, decreasing household sizes and a housing shortage of currently 20 million units will call for extensive residential construction. In addition, affordable housing as well as mid stage luxury housing is driving growth in Metro as well as tier II and tier III cities. The Indian economy is steadily moving forward on its path to prosperity with economic development being the focal point of the progress. In the post liberalization era, India has attracted huge quantum of foreign direct investment on account of its excellent economic performance and recently real estate sector has also been deregulated and liberalized. Today India is seen as a prime destination for investment by overseas investors across the board. India's favourable demographic and economic scenario makes it an attractive destination for the real estate investors.
DLF Limited
DLF Limited with a 62 years track record is India's largest real estate company in terms of revenues, earnings, market capitalisation and developable area. It has approximately 238 msf of completed development and 423 msf of planned projects with pan India presence across 30 cities. It's primary business is development of residential, commercial and retail properties. Its exposure across businesses, segments and geographies, mitigates any down-cycles in the market. DLF has also forayed into infrastructure, SEZ and hotel businesses. The Homes business caters to 3 segments of the residential market - Super Luxury, Luxury and Mid-Income. The product offering involves a wide range of products including
condominiums, duplexes, row houses and apartments of varying sizes. DLF has 216 msf of developed area under homes and residential plots.
References
1. Euromonitor.com: India statistics 2. CB Richard Ellis: market View India Office, Second & Third Quarter 2008 3. DTZ Insight: Gurgaon & Noida Offices; Growing Convergence between Markets 4. DLF Report: By Anand Rathi 5. DLF Annual Report: 2009 6. DLF City Phase 1: Property price Trend from magicbricks.com 7. Demographic Profile and Settlement Pattern of NCR: IMRB International 8. Various analyst reports and presentation by DLF to analysts