Docstoc

Superstars-Baruch-4-11-2006-cm1

Document Sample
Superstars-Baruch-4-11-2006-cm1 Powered By Docstoc
					Superstar Cities



           Joe Gyourko (Wharton)
    Chris Mayer (Columbia Business School
                  & NBER)
       Todd Sinai (Wharton and NBER)

                                            1
Introduction

   National incomes have become increasingly
    skewed after WWII (Saez, Kopczuk&Saez)
   Families have become increasingly
    segregated by income across MSAs
   House prices have grown much faster than
    average in the most attractive MSAs
   Similar pattern of right-skewness in income
    and house prices for places (roughly
    equivalent to suburbs) within MSAs
Introduction
   Increasing number of rich competing to live in
    scarce number of desirable cities
   US population grew 130 mm from 1950-2000
     San Francisco

           Absorbed only 650,000 people
           Population growth slowed as house price growth
            rates rose from 1980-2000
       Las Vegas & Houston
           Added 3.3 mm & 1.5 mm people, respectively
           Population grew most rapidly in the last two
            decades as house price growth slowed
Table 1: Real annual house price growth, top 10 MSAs
                                                          Annualized growth
      Rank                       MSA                       rate, 1950-2000
        1                 San Francisco                         3.53
        2                     Oakland                           2.82
        3                      Seattle                          2.74
        4                  Los Angeles                          2.46
        5                     Portland                          2.36
        6                      Boston                           2.30
        7             Bergen-Passaic, NJ                        2.19
        8                   New Haven                           2.12
        9                     New York                          2.09
        10                     Atlanta                          2.07
                    Weighted average of all
        16                                                      1.80
                       MSAs in sample
Note: Only MSAs with 1950 population >600,000 are considered
                         Density of Mean House Values Across MSA's
                                         1950 versus 2000

3.50E-05

                                                  Skewness in Mean House Values Across
3.00E-05
                                        2.5               MSA's:1950 vs 2000
               1950 Density
                                          2
2.50E-05
                                        1.5
2.00E-05
                                          1
                                        0.5
1.50E-05
                                          0
                                              0         1       2         3        4     5
1.00E-05
                            2000 Density
5.00E-06



0.00E+00

           0    100000         200000              300000        400000        500000        600000
Growth differences are persistent

   Long-term house price growth is skewed to a
    few places
     That growth is positively correlated over 30-

      year periods (0.3)
     One-half of cities in top quartile from 1940-

      1970 are also in the top quartile from 1970-
      2000
     Decadal price growth rates are slightly

      negatively correlated
Our proposition: Superstar cities

   Some desirable (“superstar”) cities are in
    limited supply
     Few close substitutes exist among other

       cities
     New construction is difficult

   High income families outbid others and are
    over-represented in these superstar cities
Our proposition: Superstar cities

   City price growth is driven by growth in the
    right tail of the national income distribution
     As the population grows, prices rise for the

       top cities
     As the number of high income families

       grow, rich people increasingly populate the
       superstar cities, paying higher prices
   Superstar cities need not be more productive
   Superstar cities need not be more desirable
Implications of Superstar cities

   House price growth might not reflect
    increases in underlying economic value
     Rather, it is an outcome of national growth

      in population or number of wealthy
     Certain cities/places are luxury goods

   Agglomeration might not cause productivity;
    rather, productive (high income) people
    agglomerate
Implications of Superstar cities
   House prices across cities can continue to
    diverge in the long-run
     Some cities can evolve into concentrations

      of increasingly high-income families
   Price/rent is higher in superstar cities
   House prices in superstar cities are more
    volatile (Himmelberg, Mayer, Sinai 2005)
   Low income workers may not be able to afford
    their houses and eventually move out of
    superstar cities while high income workers
    move-in
Strategy

   Use a simple model to illustrate that inelastic
    supply plus growth population can generate
    the stylized facts we have observed
   Model predictions:
     Price-rent ratio is higher in superstar cities

     Income sorting across cities

   Examine model’s predictions in three
    samples:
     Across MSAs

     At the place level (within MSA)

     Among movers
Our goals

   We demonstrate that this is a reasonable
    potential explanation for the stylized facts
     Simple mechanism with few assumptions

     Fits neatly with the evolution of MSA house

      prices and incomes
   We will not disprove other reasonable
    alternatives (many are complementary)
   We hope to show that our story must be at
    least part of the explanation
   We think the data show that it is a big part!
Model Overview

   Construct a simple model where:
     Workers live in one of two cities and also

      consume a composite good
     There is a distribution of incomes

     The intensity of preferences between

      residing in the two cities varies randomly
      across individuals
     All housing is rented for a single period
Model Overview

   Solve for an equilibrium
     Workers are freely mobile

     Rents adjust to equate supply and demand

      for living in each city
   Extend the model to allow changes/growth
Labor Market and Urban Setting

   Two cities
     Red city, R, has a fixed limit on
      development of up to K units (e.g., San
      Francisco)
     Green city, G, has a perfectly elastic supply
      of new housing (e.g, Las Vegas)
   Land rent to live in each city is:
     Land is free in G (r = 0)

     If demand is high enough, workers pay r ≥ 0
      to live in R where r is competitively
      determined
Preferences for R vs. G

   Worker i draws a preference ci to live in R vs.
    G
     ci represents the relative match quality that

      a worker receives in R vs. G
     ci ~u[0,1], where the higher the value of ci

      the stronger the preference to live in G
     Assume that ci is independent of the worker

      type
Utility Function

   Workers choose two items:
     Live in R or G

           Workers always consume one unit of land in
            either city
     A units of a composite good
   Workers get utility:
     (1-c)*A if live in R

     c*A if live in G
Equilibrium

   A worker with a preference ci is indifferent
    between living in R and G if the utility of living
    in R equals the utility of living in G
     Cutoff rent, r, is set so that R is full

     Worker trades off the rental cost of living in

      R against her preferences for being there
     Any worker with c≥0.5 lives in G at r=0

   Some workers with c<0.5 prefer R but don’t
    want to pay the equilibrium rent, r, so live in G
    anyway.
Cross-sectional income distribution, by city
                                       Figure 1
                       Wage Distribution in Red and Green Cities
  1400

  1200

  1000

   800
                        Live in Green City
   600
                                                                                                 30,000 workers
   400

   200
                                              Live in Red City
      0
                   0



                              0



                                         0



                                                    0



                                                               0



                                                                          0


                                                                                     00



                                                                                                 00



                                                                                                             00



                                                                                                                         00
      $0


                 00



                            00



                                       00



                                                  00



                                                             00



                                                                        00



                                                                                   ,0



                                                                                               ,0



                                                                                                           ,0



                                                                                                                       ,0
               5,



                          0,



                                     5,



                                                0,



                                                           5,



                                                                      0,


                                                                                 05



                                                                                             20



                                                                                                         35



                                                                                                                     50
            $1



                       $3



                                  $4



                                             $6



                                                        $7



                                                                   $9


                                                                              $1



                                                                                          $1



                                                                                                      $1



                                                                                                                  $1
Note: 100,000 workers; truncated normal distribution; mean=40,000; std.=40,000, w L=$0, w U =$150,000
What do we learn from the model?

   Proportional increase in national population
    leads to a higher rent (price) in R
   Thicker right tail of national wage distribution
    leads to a somewhat higher rent (price) in R
   Consistent growth in population or # rich
    causes R to have a higher growth rate of rents
    and house prices
   Higher rent (and price) growth in R causes the
    price/rent ratio to be higher in R (returns to
    owning real estate are same in both cities)
     Population growth
                                        Figure 2
                    Wages and Rents in Red City, Alternate Populations
  600
                                                                  Wage Distribution
                                                                  (N=200,000)
  500
                     Wage Distribution
                     (N=100,000)
  400

  300

  200
              rent=                             rent=
              $19,257                           $41,263
  100

     0
         $0     $15,000    $30,000 $45,000      $60,000    $75,000    $90,000 $105,000 $120,000 $135,000 $150,00


Note: truncated norm al distribution; m ean=40,000; std.=40,000, w L=$0, w U =$150,000
     Wage distribution shifts to the right
                                          Figure 3
                            Wages and Rents in Red City, Fatter Tails
  500
                  Wage Distribution
                  (std.=$40,000)
  400
                                                                         Wage Distribution
                                                                         (std.=$80,000)
  300


  200
              rent=                      rent=
              $19,257                    $23,994
  100


     0
         $0      $15,000   $30,000   $45,000    $60,000    $75,000    $90,000   $105,000 $120,000 $135,000 $150,00



Note: 100,000 w orkers; truncated norm al distribution; m ean=40,000; w L=$0, w U =$150,000
Alternative models

   Compensating differentials (quality of life)
     Price differences across cities due to

      differences in the capitalized value of
      amenities
     Only predicts changes over time in relative

      values of cities if amenities change over
      time (e.g., air conditioning makes southern
      cities relatively more attractive)
     Unlikely to explain the data
Alternative models

   Production agglomeration
     Price differences across cities due to

      differences in city productivity
     Productivity may have grown fastest in cities

      with concentrations of high-skill workers
   Test of production agglomeration
     Should not exist within a labor market area

      (MSA) unless productivity gains are based
      on where you live (not work)
     Can still be consumption agglomeration
Alternative models

   Correlated preferences/tournaments
     Everyone has the same preferences across

      cities
     Pure sorting with richest living in most

      desirable cities
     Sub-case of superstar cities model

     Data do not support a pure sorting model
Data

   Decennial Census (1940-2000)
     High quality but low frequency

     Aggregate from the county level up to MSAs

           Metropolitan statistical area/Primary metropolitan
            statistical area (1999 definition)
           Corresponds to the labor market area
     Has population, number of housing units, and
      distributions of income and house value
   IPUMS
     Subsample of household and person-level data

      from the Census
     Use for household or person-level analyses
Variable creation

   Distributions of House Value and Income
     Create 5 income bins with endpoints in real

      (2000) dollars
     Bins defined relative to 1980 topcode

         Chosen because it was a low topcoded decade
         0-25%, 25-50%, 50-75%, 75-100%, and
          >=100% of topcode
         Calculate from actual brackets by taking a
          weighted average of overlapped bracket
          densities
Identifying a Superstar city

   Superstars have:
     High demand

     Inelastic supply

   Implication: demand growth in a Superstar city
    should show up relatively more in prices than
    in quantity
   Graph 20-year average real house price
    growth vs. unit growth, by MSA
   Use lagged 20-years data to define Superstars
   Limits regressions to data from 1970-2000
Figure 6: Real annual house price growth versus unit growth, 1980-
2000

       .06




                                          A
       .04




                                      Nassau-Suffolk County
                                                 San Jose
                                 New York Boston                                                              Wilmington
                                   San Francisco
                                                 Santa Cruz
                                                                                  C
                                 Jersey City
                                     Pittsfield                     Asheville                              Raleigh-Durham-Chapel Hill
                                  Bergen-Passic Oakland Newburgh
                                              Trenton
                                        Springfield Salinas               Seattle        Charlotte
                               Newark New Haven County            Savannah Charleston
                                     Detroit Dutchess Portland Rapids RosaJacksonville
                                                               Grand Santa
                                                                AnnMiddlesex-Somerset-Hunterdon Atlanta
                                                  Santa Barbara-Santa Maria Athens
       .02




                                                                     Arbor
                                                              Monmouth-Ocean
                                        Philadelphia
                              Benton Harbor
                                            Providence                        San Nashville-Davidson
                                                                    Tuscaloosa Luis Obispo
                                      JacksonLouisville Falls Hickory    Greenville
                                                    Glens Wilmington Bellingham
                                                                        Greensboro-Winston-Salem
                                                                         Vallejo
                                                  ColumbusIndianapolis Charlottesville
                                                 New London            Portland
                                             Albany Chattanooga Knoxville
                               ScrantonChicago BirminghamBloomington Pensacola                                    Austin
                                                               Ventura Tacoma Bremerton
                                                                 Burlington
                                            KalamazooBaltimore Greeley    Augusta          Colorado Springs              Myrtle Beach
                                             Yakima
                                             AngelesOmaha MaconCounty Lake Sarasota
                                                   AllentownLynchburgDenver CityTallahassee West Palm Beach-Boca Raton
                                    Vineland-Millville-Bridgeton Richmond-Petersburgh
                                                            Memphis Fayetteville
                                                   ReadingElkhart San Diego
                                                  Cincinnati Columbus Springfield Provo-Orem
                                              Roanoke Rocky Mount Salt
                                        LosSt. Lansing FlorenceMadison                           Fayetteville
                                                  Bend
                                                   HagerstownLancaster Columbia Huntsville
                                     Flint SouthHarrisburg SalemNorfolk
                                                Louis
                                              Danville       Orange Medford
                                                             State College
                                                                 Stockton
                                                             Goldsboro Tampa-St. Petersburgh
                                                                      Mobile                                     Orlando
                                                     Atlantic CitySumter Sacramento Greenville
                                                       Joplin
                                                         Johnson
                                                       Kansas City
                                Kankakee HartfordClaire City Jonesboro
                         St.Altoona Gadsden EauLafayetteDecatur Albuquerque
                              Joseph Akron KenoshaDothanGreen Bay
                                   Gary Lewiston Jackson
                                                             Biloxi
                                                            York       Washington
                                                              Montgomery
                           Terre HauteRochesterSpokane Appleton
                                    SaginawPuebloFlorence Little Lexington-FayetteBoise Daytona Beach Phoenix
                                                   Eugene Hattiesburg
                                  Cleveland Shermon-Denison Rock Clarksville
                                                       Wayne Minneapolis-St. Paul
                                      Kokomo SheboyganFortHamilton
                                Utica-RomeAnniston
                           Sharon DaytonJanesville Rapids Smith Falls Gainesville City
                                       Syracuse Fort Des Moines Sioux
                                   Canton
                               Buffalo Milwaukee Bloomington
                                                  Cedar
                                                   Bangor Lincoln Cloud
                                                                    St.                Dallas
                                   Toledo
                                   Williamsport Amarillo Chico
                              Mansfield Evansville
                                   Erie
                               Muncie
                             Pittsburgh Monroe Waco Merced
                               Alexandria
                              Cumberland Racine
                                                      Miami
                                                     Wausau
                      Duluth SiouxLawton Owensboro Jackson ModestoLauderdale
                            Youngstown-Warren La Visalia-Tulare-PortervilleTucson
                                                   RockfordGalveston Killeen
                                 LimaCity TexarkanaCrosse
                                                       Yuba       Tyler
                                                                                        Lakeland
                                                                                San Antonio Worth
                                                                             Fort
                                                                                          Fort              Laredo
                                                                                       Reno Riverside-San Bernardino                    B   Las Vegas
                                           Springfield
                                      Dubuque Longview
                        JohnstownParkersburg Tulsa
                                     New
                              JamestownOrleans Richland Rouge
                        ElmiraBluff Falls
                          Pine Great Shreveport-Bossier                                          Brownsville         McAllen
             0




                                                   Albany Baton
                                                         Wichita
                                                   Lake Charles
                            Grand Forks Champaign San Angelo
                                   Charleston
                                 Beaumont
                                   Binghamton                       Houston
                                                           Oklahoma City
                         Peoria-Pekin Falls Lubbock Lafayette
                                                        Corpus Christi
                                   Wichita
                                   Huntington
                           Davenport Topeka           Billings Fargo Bakersfield
                  Wheeling
                 Steubenville-Weirton                   Houma FresnoEl Paso
                         Waterloo-Cedar Falls Abilene
                   Decatur


                        Enid
                                      D               Odessa
      -.02




                   0                                    .02                    .04                                                               .06
                                                housing units annual growth (1980-2000)
Identifying a Superstar City

   Superstars: Region “A”
     High price growth to unit growth ratio

           1.7 = sample 90th percentile
       Outside of minimum demand frontier
           Median price growth + unit growth, by decade
   Non-superstars: Region “B” High unit growth
    to price growth ratio (1.7)
   In-between: Region “C” (excluded in
    regressions)
   Control for low-demand: Region “D”
Identifying a Superstar City

   Analysis robust to other functional forms of
    price and unit growth
   This approach enables us to look at changing
    superstar status over time
     Moving into region A is like “filling up”

           Becoming more inelastic
       Lots of “filling up” between 1960 and 1990
       Within-MSA, over time variation in
        inelasticity
       Non-superstars can move out of region B
Figure 7: Real annual house price growth versus unit growth, 1960-
1980
     .08
     .06




                                                                       Santa Cruz


                                                                            Santa Rosa                           Orange County
                                                                                                             Ventura
                                 San Francisco
                                                                                   SanSan Jose
     .04




                                                                      Salinas          Luis Obispo
                                                           OaklandBellingham
                                                           Laredo             VallejoSan
                                                                                      Barbara-Santa Maria
                                                                               SantaEugene
                                        Los Angeles                Portland Modesto Diego
                                                             Fresno Bremerton
                                                                               Medford
                                                                Stockton Greeley
                                                                     Tacoma Salem
                                                                Seattle Fort Worth Provo-Orem
                                                          Visalia-Tulare-Porterville Riverside-San Bernardino
                                                                                 Denver
                                                                                     Richland
                                                                                                    Houston
                                                                                                   Austin    Phoenix
                                                         LongviewMerced
                                           Enid Atlantic City
                                                          Bakersfield
                                                                Lafayette
                                            Texarkana
                                                              KnoxvilleSmith Washington
                                                   Baltimore Tulsa                      Fayetteville Tucson
                                                                                    Sacramento
                                                                     Tyler ChicoWilmington
                                                                  Fort Houma Dallas   Killeen
                                        Evansville Shermon-DenisonMcAllen
                                                                    Nashville-Davidson                            Reno                                 Las Vegas
                                                                 Burlington Boise
                                                            Galveston Hickory City
                                                                     Yuba
                            Altoona Yakima San AngeloGreenville Charleston Myrtle Beach
                                              Spokane
                                               BirminghamMiddlesex-Somerset-Hunterdon                                                  Fort Lauderdale
                                                 Hagerstown Hattiesburg City Columbia
                                                                    Oklahoma
                                              PuebloWausauClarksville Salt Lake
                        Terre Haute Corpus ChristiMinneapolis-St. Paul CityAlbuquerque Hill
     .02




                                                 La CrosseBloomington
                                                                   Charlotte
                                                         BrownsvilleNorfolk Raleigh-Durham-Chapel                           West Palm Beach-Boca Raton
                                                     YorkSioux Falls Little Rock
                                         Portland Jackson Greensboro-Winston-Salem
                                             Trenton
                                            New HavenJonesboroGreen
                                                     Lincoln                                                Springs Sarasota
                               Cumberland Springfield FargoGreenville Baton Rouge ColoradoGainesville
                                   Joplin AllentownLancaster AugustaMiami
                                    WichitaPeoria-Pekin Johnson City Bay
                                        Sheboygan Eau Claire Decatur Monmouth-Ocean
                                CharlestonDes MoinesGoldsboroAnn Arbor
                                                        Asheville San Lexington-Fayette
                                                       Chattanooga Antonio
                                 Bergen-Passic Grand ForksSpringfield Charlottesville
                       Wheeling Reading Parkersburg LynchburgJackson
                              Scranton Huntington RoanokeFlorence Biloxi Atlanta Tampa-St. Petersburgh
                          Johnstown Waterloo-CedarMountSt. Cloud College Lakeland Tallahassee
                                                    Sumter
                                                Kokomo                Pensacola
                              Newark Savannah Anniston Tuscaloosa Bloomington
                                      Waco Lake Charles Falls Richmond-Petersburgh Fayetteville
                                                   Rocky
                                     Williamsport LondonMemphis State
                         St. Joseph Philadelphia Mobile Monroe            Madison
                                         BangorNew
                                              Davenport
                        Sioux City Providence Lafayette Billings Dothan
                                       Pine Topeka Hamilton
                                     AbileneKenosha Florence
                                             Bluff
                                                 Hartford
                                        BostonOdessa Montgomery
                                                 Harrisburg                                      Orlando
                                             Kansas
                                  Wichita Danville City
                                               Dubuque
                                     Glens FallsOwensboro Albany
                                        Gadsden
                                     MilwaukeeRacine Elkhart
                                            Falls
                                       Lima Shreveport-Bossier
                                                          Macon
                                                                               Huntsville
                                                                                     Athens
                                        ToledoFalls Appleton Orleans
                                      ErieCanton WayneLansing
                                                 Fort      New
                                      Decatur Indianapolis El
                           PittsburghGreat KankakeeColumbusLawton
                                                            Champaign                             Daytona Beach
                                         Muncie Cedar Rapids Jacksonville
                                        Chicago Alexandria
                                          Beaumont
                              Steubenville-Weirton WilmingtonPaso
                                      St. Louis Janesville Lubbock
                                                   Louisville
                          Duluth Detroit Dayton     Omaha
                                        Akron Rockford Rapids
                                         Vineland-Millville-Bridgeton
                                                Flint
                                                Newburgh
                                                SaginawDutchess County
                                                  Grand
                                        Bend
                                South CincinnatiAmarillo
                                         Gary
                                     AlbanyKalamazoo
                                          Jackson
                                     Lewiston
                                   Youngstown-Warren
                          City Cleveland Nassau-Suffolk County
                   JerseyJamestown Mansfield
                              Sharon
                                   BinghamtonColumbus
                              Pittsfield Harbor
                                   Benton Rochester
                                  Springfield
                          Elmira
           0




                                          Syracuse
                         Buffalo
                         Utica-Rome


               0                          .02               .04               .06                                                                .08
                                          housing units annual growth (1960-1980)
Superstar suburbs (census places)

   Do Census-designated places within an MSA exhibit
    the same pattern as across-MSAs?
     “Fractal” version of the Superstar Cities story

           Some places are inelastically supplied
           Model says should have high price growth and pick off the
            right tail of the income distribution
       Controls for production externalities
           Should be constant in the labor market area
           Use MSA x year fixed effects
   Define superstar/non-superstar/low demand places
    analogously to MSAs
   Four decades of data yield 1990 and 2000 useable
Table 2: Ratio of real annual price
growth to housing unit growth
               1950-1970     1960-1980   1970-1990   1980-2000
Ever a “non”-superstar MSA
 25th Pctile     0.304         0.472       0.395       0.247
 50th Pctile     0.481         0.612       0.472       0.503
 75th Pctile     0.710         0.792       0.589       0.700
Always middle-range
 25th Pctile     0.552         0.654       0.328       0.466
 50th Pctile     0.746         0.838       0.604       0.889
 75th Pctile     1.034         1.133       0.929       1.368
Ever a superstar MSA
 25th Pctile     0.655         0.704       2.300       2.028
 50th Pctile     0.684         1.179       3.135       3.052
 75th Pctile     0.850         1.556       4.812       4.554
Table 2: Ratio of real annual price
growth to housing unit growth
           1950-1970 1960-1980 1970-1990 1980-2000
MSAs that switch from “non”-superstar to superstar
 25th
            0.324     0.341      2.141     2.035
  Pctile
 50th
            0.353     0.951      2.492     2.274
  Pctile
 75th
            0.397     1.007      2.571     3.663
  Pctile
Test: Is the price-rent ratio higher in Superstar
cities?
   Do Superstar cities have high price/rent
    ratios?
   Do price/rent ratios go up when cities
    transition into superstar status?
   If so, expected rent growth must be higher
    relative to other high demand cities
   Also examine same regressions for places
    (suburbs) within MSAs
   Test: Is the price-rent ratio higher in Superstar
   cities and places?
      Estimate:
     Average Price it 
log                                    + 2                  + 3
                        = β(Superstari ) β(Non - superstari ) β(Low demandi )
                            1
     Average Rent it 
                         + 5                + 6                  +
     + β(Superstarit ) β(Non - superstarit ) β(Low demandit ) yeart + ε it
        4




      Allow fixed and time-varying superstar status
      Results are robust to other specifications
        MSA fixed effects (only time variation left)

        Different linear time trends for city status

        Year*(city status) dummies
Table 3: Dep Var: Price/rent ratio
                        MSA × year             Place × year
    Unit of Obs.        (1970-2000)            (1990-2000)
                       0.152     0.045       0.186        0.221
    Superstari
                      (0.013)   (0.013)     (0.040)      (0.021)
                      -0.022    -0.008      0.015        -0.037
    Non-superstari
                      (0.011)   (0.012)     (0.027)      (0.018)
                      -0.167    -0.139      -0.100       -0.198
    Low Demandi
                      (0.014)   (0.014)     (0.022)      (0.017)
                                0.256       0.005
    Superstarit
                                (0.024)     (0.042)
                                -0.129      -0.122
    Non-superstarit
                                (0.016)     (0.300)
                                -0.110      -0.170
    Low Demandit
                                (0.011)     (0.023)
    Fixed effects      Year      Year     MSA × year   MSA × year
Test: Income sorting

   Superstar cities should have an income
    distribution that is shifted to the right
   As a city “fills up”, rightward-shift should
    accelerate
     Due to growth in the right tail of the income

      distribution, this should increase over time
   Strategy
     San Francisco/Las Vegas example

     Regression analysis
Right tail of the income distribution is getting
thicker
                                            Change in U.S. Income Distribution, 1960-2000
                                      50

                                      45

                                      40
            Share of U.S. Income in
             Percentile Category




                                      35

                                      30

                                      25

                                      20

                                      15

                                      10

                                       5

                                       0
                                           1960         1970          1980          1990          2000

                      10 to 5%             5 to 1%   1 to 0.5%   0.5 to 0.1%   0.1 to 0.01%   Top 0.01%

Source: Saez, Emmanuel. “Reported Incomes and Marginal Tax Rates, 1960-2000:
Evidence and Policy Implications,” Table B1. NBER Working Paper 10273. January, 2004.
                                                Number and share of rich people grew in the
                                                US from 1960-2000
                                                    All MSA Income Distribution, 1980 topcode
Percent of All MSA Population In Income




                                          100


                                           90


                                           80
            Category ($2000)




                                           70


                                           60
                                                                                                                                                     All MSA Population by Income Category, 1980 topcode
                                           50


                                           40
                                                                                                                                                70




                                                                                                               Population by Income Category,
                                           30
                                                                                                                                                60

                                           20




                                                                                                                     in millions ($2000)
                                                                                                                                                50
                                           10


                                            0
                                                                                                                                                40

                                                 1960        1970         1980          1990         2000
                                                                                                                                                30
                        <39,179                  39,179-78,358   78,358-117,537   117,537-156,716   156,716+

                                                                                                                                                20



                                                                                                                                                10



                                                                                                                                                0

                                                                                                                                                      1960         1970         1980          1990          2000

                                                                                                                                 <39,179               39,179-78,358   78,358-117,537   117,537-156,716    156,716+
                                   San Francisco (big price growth) gains rich,
                                   loses poor
                                    San Francisco Income Distribution, 1980 topcode

                             100


                             90
In Income Category ($2000)
 Percent of SF Population




                             80


                             70


                             60


                             50

                                                                                                                                     San Francisco Population by Income Category, 1980 topcode
                             40


                             30                                                                                                    0.45




                                                                                                  Population by Income Category,
                             20                                                                                                     0.4

                             10
                                                                                                                                   0.35




                                                                                                        in millions ($2000)
                               0
                                                                                                                                    0.3
                                    1960        1970         1980          1990         2000
                                                                                                                                   0.25
                <39,179             39,179-78,358   78,358-117,537   117,537-156,716   156,716+
                                                                                                                                    0.2


                                                                                                                                   0.15


                                                                                                                                    0.1


                                                                                                                                   0.05


                                                                                                                                     0

                                                                                                                                          1960         1970         1980         1990         2000

                                                                                                                    <39,179               39,179-78,358   78,358-117,537   117,537-156,716   156,716+
                                   Las Vegas (big unit growth), gains rich and
                                   poor, shares constant (falling versus the US)
                                      Las Vegas Income Distribution, 1980 topcode

                             100


                             90
In Income Category ($2000)
 Percent of LV Population




                             80


                             70
                                                                                                                                            Las Vegas Population by Income Category, 1980 topcode

                             60                                                                                                      0.45




                                                                                                    Population by Income Category,
                             50
                                                                                                                                      0.4

                             40
                                                                                                                                     0.35




                                                                                                          in millions ($2000)
                             30
                                                                                                                                      0.3
                             20
                                                                                                                                     0.25
                             10

                                                                                                                                      0.2
                               0

                                    1960            1970        1980          1990         2000                                      0.15


             <39,179                39,179-78,358     78,358-117,537   117,537-156,716   156,716+                                     0.1


                                                                                                                                     0.05


                                                                                                                                       0

                                                                                                                                               1960         1970         1980         1990          2000

                                                                                                                      <39,179                  39,179-78,358   78,358-117,537   117,537-156,716   156,716+
 Differences in income distributions

      Estimate how income distributions change with
        Being a “superstar” MSA

        Transitioning into the superstar region


# in Income Bin yit
                                     + 2                  + 3
                      = β(Superstari ) β(Non - superstari ) β(Low demandi )
                         1
# of Households it
         (Superstarit ) β(Non - superstarit ) β(Low demandi ) yeart + ε it
      + β4            + 5                   + 6             +


      Results are robust to other specifications, incl. MSA
       fixed effects, different linear time trends for city
       status, and year*(city status) dummies
Table 4: The income distribution in
Superstar cities, top panel (1970-2000)
                 Left-hand-side variable: Share of MSA’s families in income bin:
                              Middle-                     Middle-
                  Rich         rich         Middle         poor          Poor
Cross-section:
                   0.016       0.014         0.024        -0.013        -0.040
Superstari
                   (0.001)     (0.001)       (0.002)      (0.004)       (0.007)

                   0.005       0.003         0.001        -0.030         0.021
Non-superstari
                   (0.001)     (0.001)       (0.003)      (0.003)       (0.008)

                   -0.009      -0.008        -0.016       -0.005         0.038
Low Demandi
                   (0.002)     (0.001)       (0.003)      (0.004)       (0.008)
Adj. R2             0.367       0.574        0.318         0.188         0.158
Table 4: The income distribution in
Superstar cities, bottom panel (1970-2000)
              Left-hand-side variable: Share of MSA’s families in income bin:
               Rich      Middle-rich     Middle     Middle-poor       Poor
              0.004         0.003        0.012         -0.005        -0.014
Superstari
              (0.001)      (0.001)       (0.003)       (0.004)       (0.008)
Non-          0.006         0.005        0.003         -0.027         0.014
superstari    (0.001)      (0.001)       (0.003)       (0.004)       (0.007)
Low           -0.007        -0.007       -0.014        -0.012         0.040
Demandi       (0.001)      (0.001)       (0.003)       (0.004)       (0.008)
              0.034         0.030        0.034         -0.014        -0.085
Superstarit
              (0.003)      (0.002)       (0.007)       (0.008)       (0.015)
Non-          -0.005        -0.008       -0.009        0.006          0.015
superstarit   (0.002)      (0.002)       (0.004)       (0.005)       (0.010)
Low           -0.005        -0.005       -0.006        0.016         -0.001
Demandit      (0.001)      (0.001)       (0.003)       (0.004)       (0.007)
Table 5: The income distribution in
Superstar places, top panel (1990-2000)
             Left-hand-side variable: Share of place’s families w/ income:

               Rich     Middle-rich     Middle    Middle-poor      Poor
              0.061        0.015        0.020        -0.033       -0.045
Superstari
              (0.003)      (0.001)      (0.002       (0.003)      (0.005

Non-          -0.018       0.003        0.023        0.029        -0.037
superstari    (0.003)      (0.001)      (0.002       (0.002       (0.004

Low           -0.058       -0.029       -0.031       0.032         0.085
Demandi       (0.003)      (0.001)      (0.002       (0.002       (0.004
Adj. R2       0.208        0.237        0.118        0.101         0.186
Table 5: The income distribution in
Superstar places, bottom (1990-2000)
              Left-hand-side variable: Share of place’s families w/ income:
                Rich     Middle-rich     Middle     Middle-poor      Poor
               0.045        0.020        0.014         -0.030       -0.050
Superstari
               (0.006)      (0.003)      (0.004)       (0.006)     (0.009)
Non-           -0.008       0.003        0.014         0.017        -0.026
superstari     (0.004)      (0.002)      (0.003)       (0.004)      (0.006)
Low            -0.048       -0.020       -0.017        0.034        0.051
Demandi        (0.003)      (0.002)      (0.002)       (0.003)      (0.005)
               0.015        -0.009       -0.020        -0.004       0.019
Superstarit
               (0.007)      (0.003)      (0.004)       (0.006)      (0.009)
Non-           -0.019       -0.005       0.005         0.016        0.002
superstarit    (0.004)      (0.002)      (0.003        (0.003)      (0.006)
Low            -0.017       -0.014       -0.023        -0.001       0.055
Demandit       (0.004)      (0.002)      (0.002        (0.002)      (0.005)
How do changes in US number of rich impact
marginal home price or # MSA rich?
   Model suggests that percentage of high-
    income families in Superstar cities grows as
    number of US rich grow
   Examine this question by interacting city
    status with number of rich in US
   Easier with Superstar places,
     Interact number of rich in MSA with place

      status
     Lots of variation across MSAs & places

     Always include MSA*year fixed effects
Table 6: Changes in overall rich &
marginal house values in Superstars
 Dep. variable:           Log of the 10th Percentile House Value
Unit of observation   MSA × year      Place × year      Place × year
Superstari ×            0.115             0.065             0.090
log(Number Richkt)      (0.041)          (0.010)           (0.011)
Non-Superstari ×        0.005            -0.103            -0.047
log(Number Richkt)      (0.037)          (0.008)           (0.009)
Low Demandi ×           -0.115           -0.152            -0.106
log(Number Richkt)      (0.045)          (0.007)           (0.009)
                                        Place,            Place,
Fixed effects:        MSA, year
                                      MSA × Year        MSA × Year
N                       1,116            31,500            15,368
Rectangular
                         Yes               No                Yes
sample?
Table 6: Changes in overall rich & share
rich in Superstars
 Dep. variable:        Share of families that are in the “rich” category
 Unit of obs.          MSA × year        Place × year       Place × year
 Superstari ×             0.021             0.016              0.025
 log(Number Richkt)       (0.002)           (0.001)           (0.001)
 Non-Superstari ×         0.004             -0.001            -0.006
 log(Number Richkt)       (0.001)           (0.001)           (0.001)
 Low Demandi ×            -0.014            -0.023            -0.031
 log(Number Richkt)       (0.002)           (0.001)           (0.001)
                                           Place,            Place,
 Fixed effects:        MSA, year
                                         MSA × Year        MSA × Year
 N                        1,116             31,510            15,368
 Rectangular sample?       Yes                No                Yes
 Adj. R2                 0.1574             0.116              0.145
Migration flows

   Model suggests high-income families
    disproportionately move into superstar cities
   Look at income distribution of recent movers
     Would expect thicker right tail of movers in

      Superstar Cities
     Use IPUMS “What MSA did you live in 5
      years ago?” to indicate movers-in
         Available 1980, 1990
   Unfortunately: income measure is post-move
  Migration flows

      Regression:

  # of In-migrants yit
                              1 (Superstarit )   2 (Non-superstarit )   t   it
Total # of In-migrantsit


      Superstar in-migrant distribution is more rich,
       less poor
        Opposite for out-migrants, but not as clear

      Differences are stat significant between
       superstar & in-migrants/out-migrants and not-
       superstars & in-migrants/out-migrants
Table 7: The difference among superstar MSAs in the
income distribution of in-migrants

                     Rich      Middle-rich   Middle    Middle-poor     Poor
In-migrants:
Superstarit
                      0.045       0.061       0.099      -0.029       -0.176
[Relative to the
                     (0.015)     (0.011)     (0.024)     (0.035)     (0.044)
   income
   distribution]      [1.67]      [2.26]      [0.85]     [-0.06]     [-0.47]
Non-superstarit
                     -0.005       -0.003     -0.004      -0.015       0.026
[Relative to the
                     (0.008)     (0.005)     (0.012)     (0.018)     (0.022)
   income
   distribution]     [-0.21]     [-0.11]     [-0.03]     [-0.03]      [0.07]
Adj. R2              0.0683       0.1114     0.1371      0.0606      0.0573
(A) Superstar/Non
   different with      Yes         Yes         Yes         No         Yes
   95% confidence?
Table 7: The difference among superstar MSAs in the
income distribution of out-migrants
                                               Income Bins

                        Rich     Middle-rich    Middle       Middle-poor    Poor

Out-migrants:
Superstarit            -0.014       -0.037      -0.054          -0.013      0.044
[Relative to the       (0.020)     (0.012)      (0.027)        (0.034)     (0.034)
income distribution]   [-0.58]     [-1.37]      [-0.46]        [-0.03]      [0.12]
Non-superstarit        -0.007       -0.012       0.011          0.010       0.002
[Relative to the       (0.010)     (0.006)      (0.013)        (0.018)     (0.017)
income distribution]   [-0.29]     [-0.44]       [0.09]         [0.02]      [0.01]
Adj. R2                0.1351      0.3026       0.2641         0.1457      0.2188
(B) Superstar/Non
different with 95%       No         Yes           Yes            No          No
confidence?
Confidence level:
                        91%         99%          99%            10%         98%
(A) – (B) ≠ 0?
Conclusion

   Evidence is consistent with Superstar cities
     Superstar MSAs have:

           Higher price-rent ratios
           Higher share high-income households, fewer
            poor
           More rich moving in, poor moving out
       Results accelerate when MSAs transition
        into Superstar status (“fill-up)
       Other MSAs show opposite pattern
       Superstar places look the same as
        Superstar cities
Conclusion

   Is this production agglomeration (production
    or human capital externalities)?
     Would have to be:

           At the place level, not just MSA
           Growing over time, matching national distribution
           Really big and growing fast
           Not worth arbitraging away by moving
           Fortuitously located in inelastic MSAs
       Almost certainly agglomeration is part of the
        story, but not all of it!
Conclusion

   Future questions
     Role of consumption agglomeration

           Peers in schools, better restaurants,…
           Superstar ski resorts
       Can this process continue indefinitely?
           Will other cities enter in “product space” to
            compete away high prices from Superstar MSAs
            or Superstar suburbs
           How much is too much?
       Impact on savings and wealth across cities

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:2
posted:8/4/2011
language:English
pages:58