Coefficients and _Standard Errors_ - American Dream Coalition

Document Sample
Coefficients and _Standard Errors_ - American Dream Coalition Powered By Docstoc
					                DO BELOW MARKET
         HOUSING MANDATES WORK?
        EVIDENCE FROM CALIFORNIA
                            Based on my research with
                            Ben Powell and Tom Means
Edward Stringham, Ph.D.
Department of Economics
San Jose State University
www.sjsu.edu/stringham
Overview
   The problem of housing affordability
   Below market housing mandates as a proposed
    solution
   Assessing below market housing mandates
     How effective have they been at producing units?
     How much do below market units cost and who pays?

   Economics of below market housing mandates
   Regression results on how below market
    housing mandates affect housing prices and
    housing quantity
   Conclusion
                      www.sjsu.edu/stringham
When I went to grad school in Virginia, I lived in
in this luxury highrise with a classmate for $655
per person.
When I got a job in California, I figured I could
live an equivalent building like this one.
www.sjsu.edu/stringham
The Problem: Housing prices are very
high
 In San Francisco the Median Priced Home
  sells for $773,500
 In Santa Clara County “the suburbs”, the
  Median Priced Home sells for $689,000

             Source: DataQuick Information Systems,
                www.dqnews.com, October 18, 2007



                 www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
The Problem: Housing prices are very
high
    That means housing payments for these
     median priced homes in Santa Clara
     County are $4,355 per month, or roughly
     $52,250 per year, or $143 per day!

        (Assuming a 30 year mortgage at 6.5 percent)



                       www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
High prices preclude many from buying




                   www.sjsu.edu/stringham
Why are prices high?
 Supply has not kept up with demand
 Are we running out of land?
     Housing  is unaffordable because of zoning
      laws (Harvard/Wharton study)
     “Exclusionary” zoning laws mandate
      minimum lot sizes, minimum density, and
      other restrictions that prevent the market
      from supplying more housing
The proposed solution

   Inclusionary zoning
    A  mandatory inclusionary zoning ordinance
      as practiced in California is an affordable
      housing mandate that requires builders to
      sell a certain percentage of their homes at
      below market rates



                    www.sjsu.edu/stringham
The goals of inclusionary zoning
 The program is touted as a way to make
  housing more affordable
 The program is touted as a way to provide
  housing for all income levels, not just the
  rich
 Helps create diverse socio-economic
  communities

                  www.sjsu.edu/stringham
How inclusionary zoning
ordinances work
   Varies by city, but most California ordinances
    require 10-20 percent of new units to be sold at
    prices affordable to low income families (defined
    as a certain percentage of median income)
   For example, in Tiburon, California a low income
    family can only afford to pay $109,800 for a
    home so:
     10 percent of new homes in Tiburon must be sold at
      $109,800
     90 percent can be sold at market rates



                       www.sjsu.edu/stringham
Where do they have it?
 Most popular in California
 Also in place in New Jersey, Virginia, and
  Maryland and are being considered in
  many other places including New York,
  Los Angeles, and Chicago.




                  www.sjsu.edu/stringham
California cities with inclusionary
zoning ordinances




              www.sjsu.edu/stringham
What are the results?




            www.sjsu.edu/stringham
Examples of below market rate
developments
Examples of below market rate
developments
Examples of below market rate
developments
Examples of below market rate
developments
Examples of below market rate
developments




             www.sjsu.edu/stringham
Examples of below market rate
developments
 Looks good right?
 Many people say the programs are a
  success and should be implemented in
  more cities




                www.sjsu.edu/stringham
Assessing inclusionary zoning
 How do advocates measure success?
 What evidence do they provide that the
  ordinances are good?
 What’s the normative standard?




                 www.sjsu.edu/stringham
                                   Number of inclusionary zoning
                                        Figure 1: Number of the Bay Area
                                   ordinances inBay Area Cities With Inclusionary Zoning
                    ities with Inclusionary




                                              50
                                              45
                                              40
                                              35
                                              30
                 Zoning




                                              25
Number of Bay Area C




                                              20
                                              15
                                              10
                                               5
                                               0
                                                 70


                                                 74




                                                 84


                                                 88




                                                 98


                                                 02
                                                72


                                                76
                                                 78
                                                80
                                                 82


                                                86


                                                90
                                                 92
                                                94
                                                 96


                                                00
                                              19
                                              19
                                              19


                                              19


                                              19
                                              19
                                              19
                                              19


                                              19


                                              19
                                              19


                                              20
                                              19


                                              19




                                              19


                                              19




                                              20
Role of Economic Analysis
 Just because a policy is becoming more
  popular does not mean it is a good idea
 Hoping something is a good idea does not
  make something a good idea
 Some policies may not be the best means
  of achieving the desired ends of increasing
  housing affordability

                 www.sjsu.edu/stringham
Role of Economic Analysis
 Inclusionary zoning sounds good to many
  people, but my coauthor and I decided to
  investigate the actual results of the policy
  rather than just looking at the expressed
  intent
 What does economics have to say?




                  www.sjsu.edu/stringham
Some research questions
 Is inclusionary zoning helping increase the
  supply of affordable housing in California?
 How costly is inclusionary zoning?
 Are there any drawbacks that have not
  been considered?




                  www.sjsu.edu/stringham
First let’s compare an estimate of
housing need to how many units
inclusionary zoning produces




              www.sjsu.edu/stringham
                              0
                                  2,000
                                          4,000
                                                  6,000
                                                          8,000
                                                                  10,000
                                                                           12,000
                                                                                    14,000
         Portola Valley
                Fairfax
             Yountville
               Tiburon
         San Anselmo
        Corte Madera
              Los Altos
             Calistoga
            Mill Valley
           Sebastopol
             Larkspur
              Clayton
          San Carlos
           Los Gatos
              Benicia
                                                                                             Need by City


      Half Moon Bay
               Cotati
            Hercules
         Healdsburg
        Pleasant Hill
            Sonoma
            Danville
          Emeryville
         Menlo Park
       San Leandro
          Petaluma
           Palo Alto
outh San Francisco
           Berkeley
     East Palo Alto
          Rio Vista
          Union City
       San Rafael




  www.sjsu.edu/stringham
         Cupertino
     Rohnert Park
       Morgan Hill
        Richmond
       San Mateo
           Novato
    Mountain View
             Napa
       Sunnyvale
      Brentwood
      Pleasanton
       Livermore
           Dublin
        Fremont
     Santa Clara
     Santa Rosa
  San Francisco
                                                                                             Association of Bay Area Government’s




                           Needs
                                                                                             2001-2006 Estimated Affordable Housing




                           zoning.

                           years by
                           multiplying




                           of Bay Area
                           inclusionary




                           Governments
                           year produced




                           Determination".
                           needs according
                           average units per
                           "Affordable" units
                           produced through




                           Five year housing
                           zoning times 5.5.)
                           under inclusionary



                           to the Association
                           (Calculated for 5.5




                           "Regional Housing
How many units does inclusionary
zoning produce?




             www.sjsu.edu/stringham
                                 2,000
                                         4,000
                                                 6,000
                                                         8,000




                             0
                                                                 10,000
                                                                          12,000
                                                                                   14,000
                                                                                              D
           Portola Valley
                   Fairfax
               Y ountville
                  Tiburon
           San Anselmo
          Corte Madera
                                                                                               etermination".




                Los Altos
               Calistoga
             Mill Valley
            Sebastopol
               Larkspur
                Clayton
            San Carlos
             Los Gatos
                Benicia
       Half Moon Bay
                 Cotati
              Hercules
           Healdsburg
         Pleasant Hill
              Sonoma
                                                                                                                                                               units per year produced under inclusionary z




              Danville
           Emeryville
                                                                                            units by Bay Area city




          Menlo Park
                                                                                                                                                               "Affordable" units produced through inclusionary z




        San Leandro
           Petaluma
            Palo Alto
South San Francisco
                                                                                                                                                                                                                 oning. (C




            Berkeley
                                                                                                                                                                                                           oning times 5.5.)




      East Palo Alto
           Rio Vista
           Union City
         San Rafael
                                                                                              Five year housing needs according to the Association of Bay Area G




          Cupertino
      Rohnert Park
                                                                                                                                                                                                                                                                           Produced Under Inclusionary Zoning




        Morgan Hill
         Richmond
        San Mateo
           Novato
   Mountain View
                                                                                                                                                                overnments "R




              Napa
        Sunnyvale
       Brentwood
       Pleasanton
        Livermore
                                                                                                                                                                             egional H




            Dublin
         Fremont
      Santa Clara
     Santa Rosa
   San Francisco
                                                                                                                                                                                                                          alculated for 5.5 years by multiplying average
                                                                                            Need versus actual production of affordable
                                                                                                                                                                                      ousing Needs
    Need versus actual production of affordable
    units by Bay Area city
                                               Figure 2: Housing Needs Versus Expected Units

   Fewer than                                       Produced Under Inclusionary Zoning


                         "Affordable" units produced through inclusionary zoning. (Calculated for 5.5 years by multiplying average
    7,000 units in       units per year produced under inclusionary zoning times 5.5.)
                         Five year housing needs according to the Association of Bay Area Governments "Regional Housing Needs

    30 years             Determination".




   Only 228         14,000


                     12,000

    Annually         10,000


   Average city      8,000



    produces fewer    6,000



    than 15 per       4,000




    year after
                      2,000


                         0

    adopting a
                                              Los Altos
                                         Portola Valley
                                                Fairfax

                                         San Anselmo
                                             Yountville
                                               Tiburon
                                        Corte Madera
                                             Calistoga
                                           Mill Valley
                                          Sebastopol
                                             Larkspur
                                              Clayton
                                          San Carlos
                                           Los Gatos
                                              Benicia
                                     Half Moon Bay
                                               Cotati
                                            Hercules
                                         Healdsburg
                                            Sonoma
                                       Pleasant Hill
                                            Danville




                                         Union City
                                         Emeryville
                                        Menlo Park
                                      San Leandro
                                          Palo Alto
                                         Petaluma
                              South San Francisco
                                    East Palo Alto
                                          Berkeley
                                         Rio Vista

                                        Cupertino
                                       San Rafael
                                    Rohnert Park
                                       Richmond
                                      Morgan Hill
                                      San Mateo
                                         Novato
                                            Napa
                                      Sunnyvale
                                 Mountain View

                                     Brentwood
                                      Livermore
                                     Pleasanton
                                          Dublin
                                    Santa Clara
                                   Santa Rosa
                                       Fremont

                                 San Francisco
    program

                                www.sjsu.edu/stringham
    Production Compared to Need
                                                         4%




                                               96%
"Affordable" units produced through inclusionary zoning. (Calculated for 5.5 years by multiplying average units
per year produced under inclusionary zoning times 5.5)
Shortfall of affordable units not produced through inclusionary zoning. (Data is only for cities with inclusionary
zoning.)
                                                www.sjsu.edu/stringham
Why does inclusionary zoning do
a poor job?
 Despite its attractive sounding name,
  inclusionary zoning is nothing more than
  a price control
 If economists agree on anything, its that
  price controls (price ceilings) on housing
  reduce the quantity and/or quality of
  housing supplied


                www.sjsu.edu/stringham
Economics of Affordable Housing
Mandates
 Price ceiling on a percentage of units
 Essentially a tax on the remainder of
  units
 Increases prices for the vast majority of
  homebuyers
 Decreases quantity of housing produced




                www.sjsu.edu/stringham
Economics of Affordable Housing
Mandates
 Price ceiling on a percentage of units
 Essentially a tax on the remainder of
  units
 Increases prices for the vast majority of
  homebuyers
 Decreases quantity of housing produced




                www.sjsu.edu/stringham
Inclusionary Zoning Creates Two Markets:
     First the Price Controlled Market
 PRICE OF                                       Supply of Housing
 HOUSING



      P1


Affordability
Control
                                                 Demand for Housing

                Qs w/ control         Qd w/control   QUANTITY OF
                          Qs w/out control=          HOUSING
                          Qd w/out control
  Inclusionary Zoning Creates Two Markets:
       First the Price Controlled Market
 PRICE OF                                       Supply of Housing
 HOUSING



      P1


Affordability              Shortage
Control
                                                 Demand for Housing

                Qs w/ control         Qd w/control   QUANTITY OF
                          Qs w/out control=          HOUSING
                          Qd w/out control
Economics of Affordable Housing
Mandates
 Price ceiling on a percentage of units
 Essentially a tax on the remainder of
  units
 Increases prices for the vast majority of
  homebuyers
 Decreases quantity of housing produced




                www.sjsu.edu/stringham
Inclusionary Zoning Creates Two Markets:
     Second the “Market” Rate Units
                          Supply of Housing w/ IZ tax
PRICE OF
HOUSING                         Supply of Housing
P w/ tax
(for market
buyers)

      P1



                                 Demand for Housing

              Q w/   Q1            QUANTITY OF
              tax                  HOUSING
 Our research was the first attempt to
  quantify the cost of the program
 Without knowing the cost of a program
  policymakers have little idea whether
  better ways of helping low income
  households exists



                 www.sjsu.edu/stringham
Sample Calculations of Cost Associated with
Providing Units for “Low” Income

                                      "Low" price control         ost
                                                                 C associated with selling "Low" unit

  $800,000

  $700,000

  $600,000

  $500,000

  $400,000

  $300,000

  $200,000

  $100,000
       $0
             Alameda




                                                          Napa
                                             Marin




                                                                                                          Solano



                                                                                                                   Sonoma
                                                                    Francisco


                                                                                San Mateo



                                                                                            Santa Clara
                       Contra Costa




                                                                       San




                                                     www.sjsu.edu/stringham
                                                    $0
                                              $100,000
                                                         $200,000
                                                         $300,000
                                                                    $400,000
                                                                    $500,000
                                                                               $600,000
                                                                               $700,000
                                                                                          $800,000
                                                                                          $900,000
                                                                                                     $1,000,000
                                                                                                     $1,100,000
                                                                                                                  $1,200,000
                                                                                                                  $1,300,000
                                     Cotati
                               Emeryville
                           Rohnert Park
                             Healdsburg
                              Richmond
                                Petaluma
                             Santa Rosa
                              Santa C  lara
                                 Hercules
                          San Francisco
                          East Palo Alto
                             Sebastopol
                                Rio Vista
                            San Leandro
                                 Sonoma
                                    So. SF
                                   Novato
                                C alistoga
                                                                                                                               price controlled unit




                                    Fairfax
                            Morgan Hill
                               Yountville
                              Union C   ity
                               Sunnyvale
                                     Napa
                             Brentwood
                               Livermore
                         Mountain View




www.sjsu.edu/stringham
                              San Rafael
                                   Dublin
                          San Anselmo
                              San Mateo
                            Pleasant Hill
                          C orte Madera
                                 Fremont
                         Half Moon Bay
                                 Berkeley
                                  Benicia
                                 Larkspur
                              Pleasanton
                                  C layton
                              San C   arlos
                               C upertino
                               Los Gatos
                               Mill Valley
                                  Danville
                                Palo Alto
                                                                                                                               Average cost associated with selling each




                             Menlo Park
                                 Tiburon
                                Los Altos
                           Portola Valley
                                                             units




                                               $25,000,000
                                               $50,000,000
                                               $75,000,000
                                              $100,000,000
                                              $125,000,000
                                              $150,000,000
                                              $175,000,000
                                              $200,000,000
                                              $225,000,000
                                              $250,000,000




                                                        $0
                             Sebastopol
                            Pleasant Hill
                                 Sonoma
                               Yountville
                         Half Moon Bay
                               Emeryville
                                   Novato
                           Corte Madera
                                     Napa
                                C alistoga
                                 Tiburon
                          East Palo Alto
                                   Dublin
                             Menlo Park
                              San C   arlos
                                 Berkeley
                              San Mateo




www.sjsu.edu/stringham
                                 Larkspur
                                  C layton
                                  Danville
                          San Francisco
                                Los Altos
                             Santa Rosa
                           San Leandro
                               Livermore
                            Morgan Hill
                               C upertino
                              Pleasanton
                                Petaluma
                                                             price controlled unit times the number of




                                Palo Alto
                                                             Average cost associated with selling each




                              San Rafael
                               Mill Valley
                              Sunnyvale
Who pays for the below market
units?
   Because government does not write a
    check for the below market units, the
    affordable housing mandate is essentially
    a tax on new housing:
     There  is no free lunch here but unfortunately
      the tax is hidden
     This hidden tax must be borne by some
      combination of market rate homebuyers,
      builders, and landowners
                      www.sjsu.edu/stringham
Inclusionary Zoning Acts as a Tax
on New Homes
(Cost per BMR unit)(% BMR Units) = Tax Per Market Unit
    (% Market Units)

For example in Mill Valley one out of ten units must be
  sold at a lost of $750,000 so:

($750,000)(10%) = $83,000 Tax Per Market Rate Unit
    (90%)

In other words, in a 10 unit development the $750,000 cost
   would be spread over the 9 market rate units.
Inclusionary Zoning Acts as a Tax
on New Homes
(Cost per BMR unit)(% BMR Units) = Tax Per Market Unit
    (% Market Units)

For example in Mill Valley one out of ten units must be
  sold at a loss of $750,000 so:

($750,000)(10%) = $83,000 Tax Per Market Rate Unit
    (90%)

In other words, in a 10 unit development the $750,000 cost
   would be spread over the 9 market rate units.
Inclusionary Zoning Acts as a Tax
on New Homes
(Cost per BMR unit)(% BMR Units) = Tax Per Market Unit
    (% Market Units)

For example in Mill Valley one out of ten units must be
  sold at a loss of $750,000 so:

($750,000)(10%) = $83,000 Tax Per Market Rate Unit
    (90%)

In other words, in a 10 unit development the $750,000 cost
   would be spread over the 9 market rate units.
Inclusionary Zoning Acts as a Tax
on New Homes
What is the magnitude of the tax in San
 Francisco Bay Area cities?




                 www.sjsu.edu/stringham
                                                $20,000
                                                $40,000
                                                $60,000
                                                $80,000
                                               $100,000
                                               $120,000
                                               $140,000
                                               $160,000
                                               $180,000
                                               $200,000
                                               $220,000




                                                     $0
                                    Cotati
                             E meryville
                             R ichmond
                          Rohnert Park
                           Santa Clara
                               H ercules
                            H ealdsburg
                                      San
                          Pleasant H    ill
                                 io
                               R Vista
                          San Leandro
                                Sonoma
                                  Novato
                              Petaluma
                           Santa R    osa
                                  Fairfax
                           Morgan H      ill
                             Sunnyvale
                                    Napa
                            Brentwood
                             Livermore
                              Mountain
                            San R   afael
                         San Anselmo
                            San Mateo
                          orte
                         C Madera
                          ast
                         E Palo Alto




www.sjsu.edu/stringham
                              Y ountville
                            Sebastopol
                                  Benicia
                                   D ublin
                                So. San
                               Larkspur
                             Union C   ity
                                 C layton
                               Calistoga
                            San C   arlos
                                                          units caused by inclusionary zoning




                                Fremont
                             Los G   atos
                             Mill Valley
                                D  anville
                           Menlo Park
                            Pleasanton
                               alf
                             H Moon
                               Berkeley
                                                          Effective tax imposed on new market-rate




                              C upertino
                                  Tiburon
                              Los Altos
                               Palo Alto
                         Portola Valley
                                                                                                                                       $225,000

                                                                                                            $200,000



                                                                                      $150,000
                                                                                                 $175,000



                                                                           $125,000

                                                                $100,000



                                            $50,000
                                                      $75,000



                                  $25,000

                             $0
                    Cotati
               E meryville
               Richmond
            Rohnert Park
             Santa Clara
                Hercules
              Healdsburg
           San Francisco
            Pleasant Hill
               Rio Vista
           San Leandro
                Sonoma
                 Novato
              Petaluma
            Santa Rosa
                 Fairfax
            Morgan Hill
            Sunnyvale
                  Napa
            Brentwood
            Livermore
        Mountain View
           San Rafael
        San Anselmo
          San Mateo
                                                                                                                                                                    the Costs Assuming 50%of tax is borne by consumers
                                                                                                                       Assuming 100% of tax is borne by consumers




       Corte Madera
       East Palo Alto
            Yountville
         Sebastopol
              Benicia
               Dublin
South San Francisco




 www.sjsu.edu/stringham
            Larkspur
          Union City
             Clayton
           Calistoga
        San Carlos
           Fremont
         Los G  atos
        Mill Valley
           Danville
       Menlo Park
       Pleasanton
    Half Moon Bay
         Berkeley
        Cupertino
           Tiburon
        Los Altos
                                                                                                                                                                                                                       Assuming 84%of tax is borne by consumers




        Palo Alto
   Portola Valley
                                                                                                                                                                    Increases in Price of New Homes Caused by Inclusionary
                                                                                                                                                                    Zoning (Under Three Different Assumptions About Who Bears
Economic theory says:

 Below market housing mandates are
  essentially a tax on the remainder of
  units
 Taxes on a product makes that product
  more expensive
 Taxes and price controls reduce the
  quantity supplied
               www.sjsu.edu/stringham
In contrast California courts say:
   The program is legal because “the
    ordinance will necessarily increase the
    supply of affordable housing”
--California Courts of Appeal in Home Builders Association
   of Northern California v. City of Napa, 2001




                       www.sjsu.edu/stringham
Whose theory is right?
   What kind of research is done?




                   www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
Whose theory is right?
   “These debates, though fierce, remain
    largely theoretical due to the lack of
    empirical research.”
    --California Coalition for Rural Housing and Non-Profit Housing
    Association of Northern California (2003, p.3)




                             www.sjsu.edu/stringham
Empirical tests
 Get data including price and quantity for all
  California cities
 Get data for when cities adopted
  ordinance
 Investigate whether adopting ordinance
  affects price or quantity of housing


                  www.sjsu.edu/stringham
Description of Data
   Sources:
     Census   data for California cities for 1990 and 2000.
     RAND California Statistics for 1990 and 2000
     Construction Industry Research Board Housing
      Permit Data 1970-2005
     California Coalition for Rural Housing and Non-Profit
      Housing Association of Northern California survey on
      policy adoption dates



                        www.sjsu.edu/stringham
                                   Variable        Observations   Mean      Standard    Minimum   Maximum
                                                                            Deviation
Table 1: Summary Statistics
                               Population 2000        N=446       65,466    (197,087)    10,007   3,694,834

                               Population 1990        N=431       58,468    (187,014)    1,520    3,485,398

                               Households 2000        N=446       22,251    (68,673)     1,927    1,276,609

                               Households 1990        N=431       20,512    (66,074)      522     1,219,770

                              Housing Units 2000      N=446       23,278    (71,843)     2,069    1,337,668


                              Housing Units 1990      N=431       21,745    (70,331)      597     1,299,963


                                Density 2000          N=446        7.62      (6.06)       0.42     37.32
                                (persons/acre)

                                Density 1990          N=431        6.87      (5.88)       0.08     37.01
                                (persons/acre)

                              Median Household        N=446       52,582    (21,873)     16,151   193,157
                                Income 2000

                              Median Household        N=431       38,518    (14,543)     14,215   123,625
                                Income 1990

                               Per Capita Income      N=446       23,903    (13,041)     7,078     98,643
                                     2000

                               Per Capita Income      N=431       16,696     (8,070)     4,784     63,302
                                     1990

                              Rents/Income 2000       N=446       27.60%     (3.1%)      14.4%     50.1%

                              Rents/Income 1990       N=431       28.9%      (2.7%)      14.9%     35.1%

                              Average Home Price      N=360       300,594   (235,436)    49,151   2,253,218
                                     2000

                              Average Home Price      N=352       206,754   (112,804)    52,858   1,018,106
                                     1990
Specification of the model for price

   Suppose we only had data for one year
    and our regression was:

Yit = β0 + d1IZyrit + β1Xit + ai + vit

HousingPrice = Intercept + Policy Dummy + Controls + Error Terms,
  which includes (ai) the unobserved city specific effects.




                            www.sjsu.edu/stringham
Specification of the model for price
And suppose the regression showed:
the intercept to be β0 = $300,000
and
the policy variable coefficient to be d1=$180,000

HousingPrice= ($300,000)+($180,000)(PolicyDummy) +….

Problem:
   Unless we could get data for relevant control variables, we would be
   unable to tell if cities with the policy have $180,000 higher housing
   prices because of the policy or because of some unobserved
   differences (ai) between cities that adopted policy and those that did
   not.


                             www.sjsu.edu/stringham
Or consider the same problem for a
cross sectional regression on
housing quantity




             www.sjsu.edu/stringham
Specification of the model for
quantity
   Suppose we only had data for one year
    and our regression was:

Yit = β0 + d1IZyrit + β1Xit + ai + vit

HousingQuantity = Intercept + Policy Dummy + Controls + Error Terms, which
  includes (ai) the unobserved city specific effects.




                              www.sjsu.edu/stringham
Specification of the model for
quantity
And suppose the regression showed
the intercept to be β0 = 70,000
and
the policy variable coefficient to be d1=40,000
HousingQuanity= (70,000)+(40,000)(PolicyDummy) +….
Problem:
  Unless we could get data for relevant control variables,
  we would be unable to tell if cities with the policy have
  40,000 more units because of the policy or because of
  some unobserved differences (ai) between cities that
  adopted policy and those that did not.


                            www.sjsu.edu/stringham
Specification of the model
   With city data, so many city specific
    characteristics are unobservable and they
    will be end up in the error term (ai)




                   www.sjsu.edu/stringham
Solution
   Panel data with first difference model
     Data  from all California cities in different years
     Our approach is estimate a first difference
      model to eliminate the fixed effect
     We also specify a semi-log model, which
      makes the interpretation of the coefficient for
      the policy variable as the approximate
      percentage change due to the change in
      policy

                       www.sjsu.edu/stringham
First difference model
Start with the level model from 2000
lnYi,2000 = d0 + d1IZyri,2000 + β1Xi,2000 + vi,2000 + ai

and the level model from 1990
lnYi,1990 = d0 + d1IZyri,1990 + β1Xi,1990 + vi,1990 + ai




                            www.sjsu.edu/stringham
First difference model
Difference the two gives us:
lnYi,2000 - lnYi,1990 = d0 + (d1IZyri,2000 – d1IZyri,1990) + (β1Xi,2000
   – β1Xi,1990) + (vi,2000 – vi,1990) + ai - ai

which can be rewritten as:
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

Notice that the city specific unobserved differences (ai) that remain
  constant over time drop out, because we are comparing cities from
  1990 to themselves in 2000. For example, San Diego is still by the
  ocean and still has nice weather in both 1990 and 2000.



                            www.sjsu.edu/stringham
First difference model
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

For the regressions on price it can be interpreted as:

Change in Price = Intercept + Change in Policy Variable +
      Change in Other Control Variables + Error Term

For the regressions on quantity it can be interpreted as:

Change in Quantity = Intercept + Change in Policy Variable +
      Change in Other Control Variables + Error Term



                            www.sjsu.edu/stringham
Interpreting the first difference
model for housing price
The first difference model looks at housing price for each
  city in 1990 and housing price for each city in 2000, and
  compare the two for each city.

So if prices in City A went from
$200,000 to $300,000
and prices in City B went from
$400,000 to $600,000,
the fact that City B has better weather and starts with
   higher prices is irrelevant if we are just looking at
   housing appreciation which is 50% for both of these
   cities.

                        www.sjsu.edu/stringham
Interpreting the first difference
model for housing quantity
Or for housing quantity, the first difference model looks at
  quantity for each city in 1990 and housing price for each
  city in 2000, and compare the two for each city.

So if quantity of homes in City A went from
100,000 to 102,000
and quantity of homes in City B went from
200,000 to 204,000,
the fact that City B is more metropolitan is irrelevant if we
   are just looking at production rates which are 2 percent
   increases for both of these cities.


                         www.sjsu.edu/stringham
Interpreting the first difference
model’s policy variable coefficient
 What we are interested in is if changes in
  the policy variable (cities that adopted the
  policy) are associated with different
  percentage changes in prices or different
  percentage changes in quantity.
 Do cities adopting inclusionary zoning
  mitigate economy wide price increases or
  do they end up with even higher prices?

                  www.sjsu.edu/stringham
Interpreting the first difference
model’s policy variable (ΔIZ)
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t


   In our data we code a 0 for cities that do not
    have a below market housing mandate and 1 for
    cities that do (for each specific year)
   Since our model is a first difference model, our
    variable (ΔIZ) looks at how a cities changes from
    time period 1 to time period 2


                                 www.sjsu.edu/stringham
Possible codings for the policy variable
ΔIZ (Changes in Inclusionary Zoning)

   ΔIZ=0 if a city has no IZ in Period 1 or 2
   ΔIZ=1 if a city has no IZ in Period 1 but has IZ in Period 2
   ΔIZ=-1 if a city has IZ in Period 1 but eliminates IZ in Period 2
   ΔIZ=0 if a city has IZ in Period 1 and keeps the policy in Period 2


   The second case is most important to us. If a city’s policy
    variable (ΔIZ) is positive it indicates a city adopted
    inclusionary zoning between period 1 and period 2



                              www.sjsu.edu/stringham
Interpreting the regression’s coefficient (d1)
for the policy variable (ΔIZ) in the price
regressions
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

   The coefficient d1 estimates what percentage the
    dependent variable (price in our first set of regressions)
    changes due to the adoption of inclusionary zoning




                            www.sjsu.edu/stringham
Interpreting the regression’s coefficient (d1)
for the policy variable (ΔIZ) in the price
regressions
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

   E.g. If the regressions looking at housing prices show
    that d1 were -0.05 it would mean adopting inclusionary
    zoning decreased prices by 5%.
   E.g. If the regressions looking at housing prices show
    that d1 were 0.01 it would mean adopting inclusionary
    zoning increased prices by 1%.




                            www.sjsu.edu/stringham
Interpreting the regression’s coefficient (d1)
for the policy variable (ΔIZ) in the quantity
regressions
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

   The coefficient d1 estimates what percentage the
    dependent variable (quantity in our second set of
    regressions) changes due to the adoption of inclusionary
    zoning




                            www.sjsu.edu/stringham
Interpreting the regression’s coefficient (d1)
for the policy variable (ΔIZ) in the quantity
regressions
ln(Yi,2000/Yi,1990) = d0 + d1ΔIZyri,t+ β1ΔXi,t + vi,t

   E.g. If the regressions looking at housing quantity show
    that d1 were 0.02 it would mean adopting inclusionary
    zoning increased quantity by 2%.
   E.g. If the regressions looking at housing quantity show
    that d1 were -0.01 it would mean adopting inclusionary
    zoning decreased quantity by 1%.




                            www.sjsu.edu/stringham
How does inclusionary zoning
affect the price of housing?




             www.sjsu.edu/stringham
Table 3: Summary of Policy Coefficients from 15 Regressions on the Price of Housing
by Model and by Lag Year
Dependent Variable: ln(Price)
                                      Level models for 2000 data         First difference models (2000-
  Level models for 1990 data                                                          1990)
Policy variable    Coefficient of   Policy variable    Coefficient of   Policy variable    Coefficient of
                  Policy Variable                     Policy Variable                     Policy Variable
    iz1985             .389             Iz1995              .627           Iz95delta           .312
    iz1986             .431             Iz1996              .642           Iz96delta           .298
    iz1987             .431             Iz1997              .637           Iz97delta           .278
    iz1988             .442             Iz1998              .637           Iz98delta           .270
    iz1989             .457             Iz1999              .642           Iz99delta           .265




                                          www.sjsu.edu/stringham
Table 5: Regression Results of How Below Market Housing Mandates Affect Price of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(average price 2000/1990)

                                                                        Coefficients and    Coefficients and
               Independent Variable                                    (Standard Errors)   (Standard Errors)
                                                                               N=431            N=431
                                                                                0.001           -0.009
                        Constant
                                                                               (0.025)          (0.025)
                                                                             0.228***
                        iz95delta
                                                                              (0.038)
                                                                                               0.217***
                        iz99delta
                                                                                                (0.037)
                                                                             0.173***          0.178***
                    Median income
                                                                             (0.0126)          (0.0125)
                                                                               -0.007           -0.008
                         density
                                                                               (0.011)          (0.011)
                                                                              -0.0017           -0.00112
                       population
                                                                             (0.00661)         (0.00662)
                                                                               -0.002           -0.003
                         rent %
                                                                               (0.005)          (0.005)
                   Adj. R-Squared                                              0.4332           0.4300
Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Table 5: Regression Results of How Below Market Housing Mandates Affect Price of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(average price 2000/1990)

                                                                        Coefficients and    Coefficients and
               Independent Variable                                    (Standard Errors)   (Standard Errors)
                                                                               N=431            N=431
                                                                                0.001           -0.009
                        Constant
                                                                               (0.025)          (0.025)
                                                                             0.228***
                        iz95delta
                                                                              (0.038)
                                                                                               0.217***
                        iz99delta
                                                                                                (0.037)
                                                                             0.173***          0.178***
                    Median income
                                                                             (0.0126)          (0.0125)
                                                                               -0.007           -0.008
                         density
                                                                               (0.011)          (0.011)
                                                                              -0.0017           -0.00112
                       population
                                                                             (0.00661)         (0.00662)
                                                                               -0.002           -0.003
                         rent %
                                                                               (0.005)          (0.005)
                   Adj. R-Squared                                              0.4332           0.4300
Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Table 5: Regression Results of How Below Market Housing Mandates Affect Price of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(average price 2000/1990)

                                                                        Coefficients and    Coefficients and
               Independent Variable                                    (Standard Errors)   (Standard Errors)
                                                                               N=431            N=431
                                                                                0.001           -0.009
                        Constant
                                                                               (0.025)          (0.025)
                                                                             0.228***
                        iz95delta
                                                                              (0.038)
                                                                                               0.217***
                        iz99delta
                                                                                                (0.037)
                                                                             0.173***          0.178***
                    Median income
                                                                             (0.0126)          (0.0125)
                                                                               -0.007           -0.008
                         density
                                                                               (0.011)          (0.011)
                                                                              -0.0017           -0.00112
                       population
                                                                             (0.00661)         (0.00662)
                                                                               -0.002           -0.003
                         rent %
                                                                               (0.005)          (0.005)
                   Adj. R-Squared                                              0.4332           0.4300
Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Summary of results
   Cities that imposed inclusionary zoning on
    average increased prices by 20 percent




                   www.sjsu.edu/stringham
Economics of Affordable Housing
Mandates
 Price ceiling on a percentage of units
 Essentially a tax on the remainder of
  units
 Increases prices for the vast majority of
  homebuyers
 Decreases quantity of housing produced




                www.sjsu.edu/stringham
How does inclusionary zoning
affect the quantity of housing?




              www.sjsu.edu/stringham
Table 4: Summary of Policy Coefficients from 15 Regressions on the Quantity of Housing
by Model and by Lag Year
Dependent Variable: ln(Housing Units)
                                     Level models for 2000 data      First difference models (2000-
  Level models for 1990 data                                                      1990)
Policy variable    Coefficient of   Policy variable    Coefficient of   Policy variable    Coefficient of
                  Policy Variable                     Policy Variable                     Policy Variable
    iz1985             .777             iz1995              .665           Iz95delta           -.045
    iz1986             .751             iz1996              .614           Iz96delta           -.024
    iz1987             .751             iz1997              .585           Iz97delta           -.027
    iz1988             .679             iz1998              .585           Iz98delta           -.038
    iz1989             .653             iz1999              .618           Iz99delta           -.051




                                          www.sjsu.edu/stringham
Table 6: Regression Results of How Below Market Housing Mandates Affect Quantity of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(units 2000 - 1990)
                                                 Coefficients and           Coefficients and
           Independent Variable                 (Standard Errors)          (Standard Errors)
                                                                               N=431      N=431
                                                                             -0.056**    -0.054**
                        Constant
                                                                              (0.023)     (0.023)
                                                                             -0.104**
                        iz95delta
                                                                              (0.042)
                                                                                         -0.097**
                        iz99delta
                                                                                          (0.041)
                                                                            0.0683***    0.0660***
                    Median income
                                                                             (0.0132)     (0.0131)
                                                                               0.113*     0.114*
                         Density
                                                                               (0.011)    (0.011)
                                                                              0.0233*     0.0230*
                       Population
                                                                             (0.00729)   (0.00729)
                   Adj. R-Squared                                              0.2921     0.2911

Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Table 6: Regression Results of How Below Market Housing Mandates Affect Quantity of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(units 2000 - 1990)
                                                 Coefficients and           Coefficients and
           Independent Variable                 (Standard Errors)          (Standard Errors)
                                                                               N=431      N=431
                                                                             -0.056**    -0.054**
                        Constant
                                                                              (0.023)     (0.023)
                                                                             -0.104**
                        iz95delta
                                                                              (0.042)
                                                                                         -0.097**
                        iz99delta
                                                                                          (0.041)
                                                                            0.0683***    0.0660***
                    Median income
                                                                             (0.0132)     (0.0131)
                                                                               0.113*     0.114*
                         Density
                                                                               (0.011)    (0.011)
                                                                              0.0233*     0.0230*
                       Population
                                                                             (0.00729)   (0.00729)
                   Adj. R-Squared                                              0.2921     0.2911

Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Table 6: Regression Results of How Below Market Housing Mandates Affect Quantity of Housing:
First Difference Model with Control Variables
Dependent Variable: ln(units 2000 - 1990)
                                                 Coefficients and           Coefficients and
           Independent Variable                 (Standard Errors)          (Standard Errors)
                                                                               N=431      N=431
                                                                             -0.056**    -0.054**
                        Constant
                                                                              (0.023)     (0.023)
                                                                             -0.104**
                        iz95delta
                                                                              (0.042)
                                                                                         -0.097**
                        iz99delta
                                                                                          (0.041)
                                                                            0.0683***    0.0660***
                    Median income
                                                                             (0.0132)     (0.0131)
                                                                               0.113*     0.114*
                         Density
                                                                               (0.011)    (0.011)
                                                                              0.0233*     0.0230*
                       Population
                                                                             (0.00729)   (0.00729)
                   Adj. R-Squared                                              0.2921     0.2911

Note: *, **,*** denotes significance at the .10, .05, .01 levels, two-tailed test.
Summary of results
   Cities that imposed inclusionary zoning on
    average decreased quantity by 10 percent




                   www.sjsu.edu/stringham
Economics of Affordable Housing
Mandates
 Price ceiling on a percentage of units
 Essentially a tax on the remainder of
  units
 Increases prices for the vast majority of
  homebuyers
 Decreases quantity of housing produced




                www.sjsu.edu/stringham
Bottom line
   Inclusionary zoning actually reduces the
    amount of housing and makes housing
    less affordable




                   www.sjsu.edu/stringham
Inclusionary zoning does not make
housing more affordable
 Inclusionary zoning is counterproductive
 Is there any silver lining?




                 www.sjsu.edu/stringham
Do price controls have any silver
lining?
   Our report has been moderately successful at
    putting some constraints on the claims by those
    who advocate inclusionary zoning
   Advocates of price controls no longer claim
    inclusionary zoning is a full solution as they used
    to, but those who still advocate the ordinance
    claim it’s a partial solution.
   They say that producing a few units is better
    than none
   They say at least it can benefit a few people
                       www.sjsu.edu/stringham
Do price controls have any silver
lining?
   Inclusionary zoning increases prices for
    most people, but could it at least benefit
    me?




                    www.sjsu.edu/stringham
www.sjsu.edu/stringham
www.sjsu.edu/stringham
$300,000
Looks like a great deal right?




              www.sjsu.edu/stringham
$300,000
Looks like a great deal right?
   What advocates of inclusionary zoning often fail
    to tell people:
     this  San Francisco condo will have resale price
      restrictions for the next 55 years
     I am age 32, so that means I would not be able to sell
      it at market rate until I am age 87
     Meanwhile, the already higher market rate units
      appreciate at normal rates creating further disparity
      between neighbors


                        www.sjsu.edu/stringham
$300,000
Looks like a great deal right?
   What advocates of inclusionary zoning often fail
    to tell people:
     this  San Francisco condo will have resale price
      restrictions for the next 55 years
     I am age 32, so that means I would not be able to sell
      it at market rate until I am age 87
     Meanwhile, the already higher market rate units
      appreciate at normal rates creating further disparity
      between neighbors


                        www.sjsu.edu/stringham
                     Appreciation of Housing Under Inclusionary Zoning
                   (Assuming market rate homes appreciate at 3% per year)


               $4,500,000
               $4,000,000
               $3,500,000                                                                          Price Controlled
                                                                                                   Home
Resale Price




               $3,000,000
               $2,500,000                                                                          Market Rate Home
               $2,000,000
               $1,500,000
               $1,000,000
                $500,000
                       $0
                            2005
                                   2011
                                          2017
                                                 2023
                                                         2029
                                                                2035
                                                                       2041
                                                                              2047
                                                                                     2053
                                                                                            2059
                                                         Year
                                                        www.sjsu.edu/stringham
                     Appreciation of Housing Under Inclusionary Zoning
                   (Assuming market rate homes appreciate at 7% per year)


               $35,000,000
               $30,000,000
                                                                                                   Price Controlled
               $25,000,000                                                                         Home
Resale Price




               $20,000,000                                                                         Market Rate
                                                                                                   Home
               $15,000,000
               $10,000,000
                $5,000,000
                       $0
                             2005
                                    2011
                                           2017
                                                  2023
                                                         2029
                                                                2035
                                                                       2041
                                                                              2047
                                                                                     2053
                                                         Year                               2059
                                                    www.sjsu.edu/stringham
Other questions aspects of
affordable housing mandates
   Is it really ownership if a person cannot gain any
    appreciation?
   Is it really ownership if a person cannot give their
    home to their children unless their children are
    also low income?
   Is a program that creates two tiers of ownership
    really good for low income families?
   How costly are these programs to monitor?
   What will be the long run effects?


                       www.sjsu.edu/stringham
What do “owners” of these price
controlled units have to say?




             www.sjsu.edu/stringham
How Should We Deal With High
Prices?
 Worst Idea….price controls
Inclusionary zoning has many problems that
  will only get worse over time
Inclusionary zoning does not address the
  real reason why housing has become so
  unaffordable


                www.sjsu.edu/stringham
How Should We Deal With High
Prices?
 Worst Idea….price controls!
 Inclusionary zoning has many problems
  that will only get worse over time
 Inclusionary zoning does not address the
  real reason why housing has become so
  unaffordable


                 www.sjsu.edu/stringham
      Just say no to price controls!

 PRICE OF                                       Supply of Housing
 HOUSING



      P1


Affordability
Control
                                                 Demand for Housing

                Qs w/ control         Qd w/control   QUANTITY OF
                          Qs w/out control=          HOUSING
                          Qd w/out control
As an alternative to price controls how
can we encourage more affordable
housing?
   Better idea…




                   www.sjsu.edu/stringham
Allowing supply to keep up with demand

PRICE OF                           Supply of Housing 1
HOUSING



     P1




                                    Demand for Housing

                        Q1               QUANTITY OF
                                         HOUSING
Allowing supply to keep up with demand

PRICE OF                          Supply of Housing 1
HOUSING


                                     Supply of Housing 2
     P1


     P1

                                    Demand for Housing


                      Q1     Q2          QUANTITY OF
                                         HOUSING
Real Solutions
(as alternatives to price controls)

   Eliminate Exclusionary Zoning,
    Eliminate Growth Boundaries, Eliminate
    Permit Moratoria, and Eliminate
    Inclusionary Zoning….



                  www.sjsu.edu/stringham
My favorite quote on this subject

 “Production is the key for being able to
  have a wide range of housing options,”
  said Michael Houlemard, executive
  director of the Fort Ord Reuse Authority. “If
  we encourage production….that alone is
  going to either stabilize or drive down
  home prices in the area.”
 (The Californian, Salinas, CA, January 19,
  2004)
                  www.sjsu.edu/stringham
My favorite quote on this subject

   “Houlemard draws his assessment directly
    from a study done by two San Jose State
    economists.”
     (The Californian, Salinas, CA, January 19,
    2004)




                   www.sjsu.edu/stringham
                DO BELOW MARKET
         HOUSING MANDATES WORK?
        EVIDENCE FROM CALIFORNIA
                            Based on my research with
                            Ben Powell and Tom Means
Edward Stringham, Ph.D.
Department of Economics
San Jose State University
www.sjsu.edu/stringham
Inclusionary Zoning Advocates
Speak
   “The price of housing is not a function of its
    development cost. Rather, housing price, be
    it rents or sale prices, are solely a function of
    market demand” (David Paul Rosen 2004).
   “Even if their profits are not maximized,
    developers will still realize acceptable profits.
    Therefore, developers will still develop”
    (Padilla 1995).
   Institute for Local Self Government states that
    inclusionary zoning helps, “Offset the demand
    on housing that is created by new
    development.” www.sjsu.edu/stringham
Advocates of inclusionary zoning
speak
   “High enough density bonuses create
    affordable units at no cost to landowners,
    developers, or other homeowners” (Padilla
    1995).
   "Most inclusionary rules are actively sought
    by developers, and can hardly be considered
    taxes" Dietderich (1996).
   “Developers often fail to participate because
    they do not understand the economics of the
    program” Kautz (2002).
                    www.sjsu.edu/stringham
    Long-Term Controls
                                         $50,000
   Income                               $45,000
    Targeting                            $40,000

   Mobility                             $35,000

   Improvements
                     Income Per Person




                                         $30,000

                                                                                                      Median Income
   Administration                       $25,000
                                                                                                      Mean Income

                                         $20,000

                                         $15,000

                                         $10,000

                                          $5,000

                                             $0
                                                   15-24    25-34   35-44       45-54   55-64   65+

                                                   www.sjsu.edu/stringham Age
     Reaction to our research
   Research has been featured in over seventy papers,
    including favorable stories in San Francisco
    Chronicle, San Jose Mercury News, Sacramento
    Bee, and Miami Herald
   In the past twelve months the report has been
    downloaded from Reason’s website 73,364 times




                      www.sjsu.edu/stringham
Reaction
   The Critics:
     “Their  paper suggests that the “market” will solve our
      housing problems. Funny that it hasn’t yet!” - Gary
      Patton, LandWatch
     “It theorizes but offers no proof, that developers pass
      the costs of the IH units to market-rate consumers...
      In reality, developers are not philanthropies and will
      charge the highest price the market will bear, with or
      without IH.” - Rob Wiener, California Coalition for
      Rural Housing.

                         www.sjsu.edu/stringham
Reaction
   The Best:
     “At  best, using IZ to provide low-income housing is at
      like fighting a forest fire with a garden hose. Under
      the harsh light this new study shines on the policy that
      hose may be spraying fuel, rather than water, on the
      fire.” - Daniel Weintraub, Sac Bee.




                        www.sjsu.edu/stringham
DO AFFORDABLE HOUSING
MANDATES WORK?
Benjamin Powell and Edward Stringham




                                  Reason
                                  POLICY
                                  STUDY

         www.sjsu.edu/stringham   318

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:2/21/2013
language:English
pages:137