How Rents and Expenditures Depreciate: A Case of Tokyo Office Properties

Transcription

1 How Rents and Expenditures Depreciate: A Case of Tokyo Office Properties March 27, 2018 Hitotsubashi-RIETI Workshop on Real Estate and the Macro Economy JIRO YOSHIDA (PENN STATE & UNIV. OF TOKYO) KOHEI KAWAI (XYMAX REAL ESTATE INSTITUTE) DAVID GELTNER (MIT) CHIHIRO SHIMIZU (NIHON UNIVERSITY) (DAVID GELTNER PRESENTING TODAY)

2 Motivation Rents (& resulting net income) measure the productivity of the physical capital represented by built commercial properties Such productivity is fundamental economic reason why buildings are built Rents reflect value of the structure plus value of the location, for existing users (in existing building) Depreciation is important economic phenomenon, major impact on use of vital resources (construction), as depreciation ultimately leads to demolition & redevelopment Rent depreciation is rarely studied (worldwide). (Most studies are of structure value depreciation.) Unique, great database available from Xymax Corp in Japan Case of Tokyo office properties has global interest as an example of extremely high land value real estate From urban economics perspective, interesting comparisons possible with land rich country (relatively low land values) such as USA where recent depreciation studies of commercial properties have been made. RENT AND EXPENDITURE DEPRECIATION 2

3 The big picture Property value (including land) depreciates only down to 60% of starting value. ==> Land Value = 60%. Age profile of commercial property value in Tokyo (Yoshida, 2017) Depreciation Rates: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years How does this compare to low land value USA case? RENT AND EXPENDITURE DEPRECIATION 3

4 The big picture Property Value/Age Profile flattening at 30 years age ==> Building fully-depreciated at 30 yrs age (land does not depreciate). ==> Short avg building economic life. Depreciation Rates: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years How does this compare to low land value USA case? RENT AND EXPENDITURE DEPRECIATION 4

5 The big picture Here is corresponding USA case for commercial property. Value/Age Profile does not flatten until 100 yrs age, at 30% Land Value Fraction Ratio to New Property Value Property Value/Age Profile (including land): Non-Parametric & Geometric/Linear Fit (Based on hedonic price model of 80,431 transaction prices in property asset market) Commercial Properties: LVF = median age (23yrs) 3.1%/yr of remaining struct value (1st 50 yrs) Building Age (Yrs) Redvlpt = 100% of old 30% of new Net Depreciation (non-parametric) Net Depreciation (geometric fit) Land Value This is for commercial property, national average. RENT AND EXPENDITURE DEPRECIATION 5

6 Geltner-Bokhari (2015), First study with sufficient data to construct empirical based survival probability curve for commercial buildings in U.S. (life expectancy = 100 yrs). Corresponds to Value/Age Profile flattening. 50% Prob Surv 105 yrs Mean lifetime = 100 yrs; Median lifetime = 105 yrs.

7 How do we know relevant land value fractions (30% commercial, 20% apartments)? Two ways 1) Direct evidence from transaction prices of properties bought as development sites subsequently sold developed, Ratio of prices: Mean NDLVF Apartment Commercial N Total ) Survival analysis (ages of demolition) life expectancy (100 yrs) combined with remaining property value fraction (30%, 20%) at age where nonparametric (flexible) value/age profile flattens out ( yrs), indicating no further depreciation (just land value): 1 Estimated Non-parametric Property Value/Age Profiles: Commercial & Apartment Ratio to New Property Value Building Age (Yrs)

8 The big picture Here is picture for three different USA metros, with three different levels of land value (Bokhari-Geltner, REE 2016) Cumulative Effect of Real Depreciation on Property Value (including land): Comparison of Several Metro Areas Ratio to Zero-Age Property Value New York Chicago Dallas Property Age (yrs) NY Chi Dallas New York flattens at 60% at 75 yrs age. Dallas at < 10% over 120 yrs. RENT AND EXPENDITURE DEPRECIATION 8

9 The big picture We found three stage lifespan of USA buildings: Youth, Middle Age, Old Age Property Value/Age Profiles: Commercial & Apartment Properties (Property Asset Value Including Land) Ratio to New Property Value Youth (0-30yrs) (Loses "Class A") Middle Age (30-65yrs) ("Class B" absorbs CapEx) New York Old Age (65-100yrs) (Declines to Land Value) Chicago Dallas Commercial (80,431 obs) Building Age (Yrs) Apartment (27,374 obs) RENT AND EXPENDITURE DEPRECIATION 9

10 The big picture This seems to exist in Japanese buildings too. Rapid depreciation in Youth (But youth may be shorter.) Age profile of commercial property value in Tokyo (Yoshida, 2017) Depreciation Rates: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years RENT AND EXPENDITURE DEPRECIATION 10

11 The big picture Depreciation more broadly from economic perspective is Capital Consumption, which includes: (i) Net Depreciation, + (ii) Capital Improvement Expenditures ( Capex ) USA Annual Gross Depreciation Rates for 25-year-old Building: Commercial: Apartment: Percent of Value of: Property Structure Property Structure Net Depreciation 1.63% 3.14% 2.38% 3.94% Capex 1.75% 3.47% 1.96% 3.36% Gross Depreciation 3.39% 6.61% 4.34% 7.30% Tokyo Office Net Depreciation Rates, Fraction of Property Value Including Land: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years Depreciation rates can be measured either as fraction of total property value (including land), or as fraction just of remaining structure value (since structure is what depreciates). In the case of Rent Depreciation, we measure it as a fraction of itself (fraction of rent), but rent reflects land value (for existing bldg) as well as structure value. Rent is current productivity of the property, not Present Value of all future expected cash flows from property. RENT AND EXPENDITURE DEPRECIATION 11

12 The big picture Here is what we find about the Capex/Age Profile for Tokyo offices Capex tends to increase with building age. RENT AND EXPENDITURE DEPRECIATION 12

13 The big picture Similar finding for USA commercial property capex $4.00 $3.50 $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 NCREIF Commercial (Annual Capex/Sqft) $ Building Age (Yrs) Commercial (Capex/Sqft) Comm (Capex/Sqft) - No LCs Capex tends to increase with building age. (With or without including Leasing Commissions ) RENT AND EXPENDITURE DEPRECIATION 13

14 The big picture Back to a key question Why/How does the Value/Age Profile flatten out so early for Tokyo commercial properties? Why do Japanese buildings have such a short economic lifespan?... Age profile of commercial property value in Tokyo (Yoshida, 2017) Depreciation Rates: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years RENT AND EXPENDITURE DEPRECIATION 14

15 The big picture Is it because of decline in the productivity of the buildings? If so, this would be indicated by Rent Depreciation. Or is it because of economic obsolescence (Redevelopment Option Value)?... Age profile of commercial property value in Tokyo (Yoshida, 2017) Depreciation Rates: 5.3%/y for 1-5 years, 2.1%/y for years, 0.9%/y for years RENT AND EXPENDITURE DEPRECIATION 15

16 Newly Contracted Rents: Building Panel This is may be our single most important empirical finding. Rents in Tokyo Office Properties decline with bldg age only very little, less than decline in property value Baseline Control Cohort 1.4%/y for ages %/y for ages %/y for ages %/y for ages %/y for ages RENT AND EXPENDITURE DEPRECIATION 16

17 Rent/Age Profile nearly flat. Declines with bldg age only about 30% to about 70% of original value. Then rent flattens, even though rents do not reflect the redevelopment option value in the land. ==> Buildings maintain very high productivity at age where economic obsolescence occurs. Baseline Control Cohort 1.4%/y for ages %/y for ages %/y for ages %/y for ages %/y for ages RENT AND EXPENDITURE DEPRECIATION 17

18 This Study We do more than just study rents on new leases Estimate age profiles of complete office cash flows New rents, rents for sitting tenants, average rents, operating expenses, net operating income, capital expenditures, and net cash flow Proprietary data from Xymax, a major property management firm in Tokyo. RENT AND EXPENDITURE DEPRECIATION 18

19 Main Results 1. The rent depreciation rate is 0.8% for new leases, 0.4% for sitting tenants, and 0.5% on average 2. Rents depreciate at a declining rate. 3. Smaller buildings experience larger rent depreciation. 4. A tenant occupying a larger proportion of building experiences larger rent depreciation. 5. Operating expenses depreciate annually at 0.6%. 6. Net operating income depreciates annually at 0.4%. 7. Capital expenditures generally increase over time. 8. Net cash flows depreciates at 0.6% per year. RENT AND EXPENDITURE DEPRECIATION 19

20 Rent and Cash Flow Data Buildinglevel Annual Panel (6159 bldgs.) Unit-level Monthly Panel (162,559 obs.) New Rent Average by year & building New Rent Average Rent Average by year & building Individual Lease Rates Tenants occupying consecutive years Sitting Tenant Rent A chaintype index Individual Lease Rates (sitting tenants) Operating Expenses Net Operating Income Capital Expenses Net Cash Flow RENT AND EXPENDITURE DEPRECIATION 20

21 Descriptive Statistics: Entire building panel (bldg. & year) Variable n mean sd median min max BLDG_ID 21,415 7, , , , YEAR 21,415 2, , , , NEW_RENT 19,993 64, , , , , AVG_RENT 1,930 63, , , , , SIT_RENT 1,930 67, , , , , CAPEX 1,965 4, , , , OPEX , , , , Net Operating Income , , , , , Net Cash Flows , , , , , GFA 21,415 12, , , , NRA 21,415 7, , , , HEIGHT 21, DISTANCE 21, , MINUTES 21, AGE 21, COMPLETION_YEAR 21,415 1, , , , RENEWAL 21, RENEWAL_YEAR 4,266 2, , , , DEMOLITION 21, DEMOLITION_YEAR 561 2, , , , RENT AND EXPENDITURE DEPRECIATION 21

22 Empirical Strategy Rents, Operating Expenses, Capital Expenditures lnvv iiii = aa 0 + ff AA iiii + XX iiii bb + YY tt + CC dd + εε iiii, VV iiii : the variable of interest for building ii at time tt, Age function ff AA iiii (1) the linear model, ff AA iiii = aa 1 AA iiii (2) the spline function, ff AA iiii = TT nn=0 aa 1,nn II nn, where II nn is the age dummy. YY tt : year fixed effects CC dd : decennial cohort (vintage) effects XX iiii : log gross floor area, walk minutes from station, city, past renovation Cash Flows (NOI, NCF) CCCC iiii = aa 0 + ff AA iiii + XX iiii bb + YY tt + εε iiii RENT AND EXPENDITURE DEPRECIATION 22

23 Result RENT AND EXPENDITURE DEPRECIATION 23

24 Newly Contracted Rents: Building Panel LHS: log new rents (1) (2) (3) (4) GFA_LOG *** *** *** *** MINUTES *** *** *** *** RNW *** *** *** *** AGE *** *** _cons *** *** *** *** AgeFE(1yr) No Yes No Yes CohortFE No No Yes Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 24

25 Newly Contracted Rents: Building Panel Baseline Control Cohort 1.4%/y for ages %/y for ages %/y for ages %/y for ages %/y for ages RENT AND EXPENDITURE DEPRECIATION 25

26 Newly Contracted Rents: Large/Small Buildings LHS: log (1) (2) (3) (4) new rents Large25% Large25% Small25% Small25% GFA_LOG *** *** *** *** MINUTES *** *** *** *** RNW *** *** * * AGE *** *** _cons *** *** *** *** AgeFE(1yr) No Yes No Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 26

27 Newly Contracted Rents: Large/Small Buildings Large Small RENT AND EXPENDITURE DEPRECIATION 27

28 Rents for Sitting Tenants LHS: log sitting tenant rents (1) (2) (3) (4) GFA_LOG *** *** *** *** MINUTES *** * * * RNW AGE ** _cons *** *** *** *** AgeFE(1yr) No Yes No Yes CohortFE No No Yes Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 28

29 Sitting Tenant's Rent Baseline Control Cohort RENT AND EXPENDITURE DEPRECIATION 29

30 New and Sitting Tenant Rents differently respond to market conditions 0.5 Year fixed effects (Hedonic Price Index) New Leases Sitting Tenants RENT AND EXPENDITURE DEPRECIATION 30

31 Stickiness in Sitting Tenants Rents. Tenants have less holdup threat than landlords, less negotiating leverage? Why?... (& bldg may retain greater functionality for sitting tenants) 0.5 Year fixed effects (Hedonic Price Index) New Leases Sitting Tenants RENT AND EXPENDITURE DEPRECIATION 31

32 Average Rents LHS: log average rents (1) (2) (3) (4) GFA_LOG *** *** *** *** MINUTES *** *** *** *** RNW AGE *** _cons *** *** *** *** AgeFE(1yr) No Yes No Yes CohortFE No No Yes Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 32

33 Average Rents Baseline Control Cohort RENT AND EXPENDITURE DEPRECIATION 33

34 Average Rents: Large/Small Buildings (1) (2) (3) (4) Large25% Large25% Small25% Small25% GFA_LOG ** ** MINUTES *** *** RNW * AGE * _cons *** *** *** *** AgeFE(1yr) No Yes No Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 34

35 Average Rents: Tenants with Large/Small Presence (1) (2) (3) (4) Large25% Large25% Small25% Small25% GFA_LOG *** *** *** *** MINUTES * * * ** RNW AGE *** _cons *** *** *** *** AgeFE(1yr) No Yes No Yes YearFE Yes Yes Yes Yes CityFE Yes Yes Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 35

36 Operating Expenses LHS: log operating expenses (1) (2) GFA_LOG MINUTES RNW AGE _cons *** *** AgeFE(1yr) No Yes YearFE Yes Yes CityFE Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 36

37 Operating Expenses RENT AND EXPENDITURE DEPRECIATION 37

38 Net Operating Income LHS: net operating income (1) (2) GFA_LOG ** ** MINUTES * RNW AGE _cons AgeFE(1yr) No Yes YearFE Yes Yes CityFE Yes Yes N adj. R Rate 0.4%/year RENT AND EXPENDITURE DEPRECIATION 38

39 Net Operating Income RENT AND EXPENDITURE DEPRECIATION 39

40 Capital Expenditures LHS: log capital expenditure (1) (2) GFA_LOG MINUTES RNW AGE *** _cons *** *** AgeFE(5yr) No Yes YearFE Yes Yes CityFE Yes Yes N adj. R RENT AND EXPENDITURE DEPRECIATION 40

41 Capital Expenditures RENT AND EXPENDITURE DEPRECIATION 41

42 Net Cash Flows LHS: net cash flow (1) (2) GFA_LOG ** ** MINUTES RNW AGE * _cons AgeFE(5yr) No Yes YearFE Yes Yes CityFE Yes Yes N adj. R Rate 0.6%/year RENT AND EXPENDITURE DEPRECIATION 42

43 This is bottom line of property current productivity with existing building. Profile almost flat. Maybe even evidence rises toward end (40-50 yrs age). Suggests economic ( external ) obsolescence is major reason for demolition & redevelopment in Tokyo. Net Cash Flows RENT AND EXPENDITURE DEPRECIATION 43

44 Conclusion Cash Flow (Building Productivity) depreciation rate is very small (0.4%-0.6%/year) This rate is less than property value depreciation (Yoshida 2017) This is the first study to uncover this point. Suggests Tokyo buildings demolished (redeveloped) when building still highly productive (economic obsolescence: redevelopment option value). But why so much more so in Japan compared to USA. Can land value difference fully explain it?... We plan to conduct additional analysis by introducing rent uncertainty, etc. RENT AND EXPENDITURE DEPRECIATION 44