A Quarterly Review oBusiness andEconomic Conditions V. 19, N. 3 J2011 The Feder al reserve Bank oF sT. louis CenTral t ameriCa’s eConomy ® Frecsures Te Roles oPredatory Lending and Household Overreaching Hspncs Populatio n Growt h in District and in Nation Commodit Price GainsSpecut n vs. Fundment s
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Commodit Price GainsBy Brett Fawley and Luciana Juvenal
Commodities o all sorts have risen in price over the past ewyears. Some say that the prices re ect a bubble, driven by lowinterest rates and excessive speculation. Others say the price gains
can be ully explained by supply and demand. Is either right?
4The RegionalEconomist
July 2011 | VoL. 19, no. 3
3 p r e s i d e n t ’s m e s s a g e
10 The Mismatch between Job Openings, Job See ers
By Maria E. Canonand Mingyu Chen
oday’s high unemployment rateis o en linked to a structuralimbalance—a mismatch betweenthe skills and location requiredto ll vacant jobs and the skillsand geographical pre erences o the unemployed. But the evi-dence downplays the role o this mismatch.
12 The Forec os re CrisisBy William R. Emmons, KathyFogel, Wayne Y. Lee, Liping Ma,Deena Rorie and imothy J. Yeager
At least early in the nancialcrisis, the high rate o oreclosures
seemed to be due more to house-holds’ overreaching than to preda-tory lending. A disproportionatenumber o those being oreclosedon were well-educated, well-o ,relatively young people.
16 A C oser loo
at Ho se Price IndexesBy Bryan Noethand Rajdeep Sengupta
racking house prices is o increasing importance to many people. Tere are several pro-minent house price indexes orthe U.S. Knowing how they di er can help people decidewhich one to ollow.
1 8 c o m m u n i t y p r o i L e
D Q oin, I .By Susan C. Tomson
Tis Southern Illinois city aimsto diversi y beyond the state air
or which it’s best known. Oneapproach has been to nd uses
or the extensive airgrounds orthe other 355 days o the year.
2 1 n a t i o n a L o Ve r Vi e w
Recover Contin esBy Kevin L. Kliesen
Although the pace o economicactivity has been inconsistent andsomewhat lackluster, the overalleconomic environment isexpected to keep improving.
2 2 d i s t r i c t o Ve r Vi e w
Examining the Growtho Hispanic Pop ationBy Rubén Hernández-Murilloand Christopher J. Martinek
Between 2000 and 2010, Hispa
accounted or more than hal growth in total U.S. populationIn the Eighth District, the role Hispanics’ growth was much ledramatic—except in rural area
24 Segregation IndexShows Dec ineBy Alejandro Badel and Christopher J. Martinek
Te Index o Dissimilarity suggests that segregationdeclined or all our majormetropolitan areas in the EightDistrict between 1970 and 200A breakdown o the index helpto show how this happened.
26 economy at a gLa nce
27 reader exchange
The Regional Economist blq l b r p bla lr v B k s . L . i
l, l l , l l
l el r v d . V
l s . L l
r v s .pl
s b B 314-444-7425 b - l b .b @ l . b. . y
l b l .s b l
v b / bl The Regional
Economist l . w v
l l l .
Director of Researchc J. w ll
Senior Policy Advisercl c. c l
Deputy Director of Researchd v c. w l kDirector of Public AffairsK B
Editors b B
Managing Editoral s b k
Art DirectorJ w ll
s l - b .t b b , - l l. .@ l . b. v .
l . / bl . y l The Regional Economist ,
p bl a o , l r vB k s . L , B 442, s . L ,mo 63166.
he ederal unds rate has been close tozero since December 2008, when the
Federal Open Market Committee (FOMC) voted to reduce the target to between0 percent and 0.25 percent. With its policy rate near the zero bound, the FOMC turnedto large-scale asset purchases (so-calledquantitative easing) as economic conditionswarranted urther action. Quantitative eas-ing was success ul and showed that the Fedcan conduct e ective policy even with the
ed unds rate near zero.Te FOMC’s rst quantitative easing
program, which began in late 2008 andended in the rst quarter o 2010, consistedo purchases o agency debt, agency mortgage-backed securities and longer-term reasury securities. Te program isgenerally considered to have been success-
ul in urther easing monetary conditions.Troughout the spring o 2010, however,
nancial market stress in the U.S. increasedagain, mostly in response to an intensi ca-
tion o the European sovereign debt crisis.During the summer o 2010, the pace
o the U.S. economic recovery slowed. Inaddition, in ation and expected in ationwere both quite low—some measures wereas low as they had been in 50 years. In a-tion, while still positive, had been trendingdownward (which is known as disin ation)throughout the rst hal o 2010. As theJapanese experience over the past 15 yearshas shown, having mild de ation (i.e., declin-ing prices) along with a near-zero policy rate
can lead to poor economic outcomes, and thesituation is di cult to escape. 1, 2 Avoiding asimilar experience in the U.S. was one o theprimary motivations or a second round o quantitative easing.
Fed Chairman Ben Bernanke gave a speechin Jackson Hole, Wyo., on Aug. 27, 2010, inwhich he rst indicated that a second asset-purchase program may be needed. At theNov. 2-3 meeting, the FOMC made the deci-sion to purchase reasury securities at a pace
o about $75 billion per month through therst hal o 2011 or a total o $600 billion—
the program commonly known as QE2. 3
Te policy change was largely priced intothe markets ahead o the November FOMCmeeting, as nancial markets are orward-looking. Te nancial market e ects o QE2were entirely conventional. In particular,real interest rates declined, expected in ationincreased, the dollar depreciated and equity prices rose. Te purchases o longer-term
reasury securities essentially lowered therisk- ree real interest rate, which then causedsome investors to switch to riskier assets—most notably U.S. equity markets, but alsoemerging market equities and commoditiesas an investment class—in search o higherrates o return.
Following the November decision, many people expected the program to have noimpact. Some even went so ar as to say that purchasing $7 trillion in longer-termbonds was necessary. But based on the
airly substantial nancial market impacto $600 billion in purchases, those viewshave been dispelled.
While the e ects on nancial marketsoccurred during the run-up to the Novemberdecision, e ects on the real economy (e.g.,consumption and employment) are expectedto occur six to 18 months a er the monetary policy action, as is the case with conven-tional monetary policy. Determining exactly which movements in real variables are due tomonetary policy and which ones are due to
other in uences on the economy that occurin the meantime can be di cult. Disentan-gling these e ects is a standard problem inmonetary policy analysis. However, the reale ects o the asset-purchase program willmost likely be conventional, just as the nan-cial market e ects were.
As the experience with quantitative easinghas shown, monetary policy can be e ectiveeven when nominal interest rates are at thezero bound. QE2 was success ul as a classic
James B ard , p ceol r v B k s . L
Te Efectiveness o QE2
p r e s i d e n t ’ s m e s s a g e
E N D N O E S1 For more discussion, see Bullard, James. “Seven
Faces o ‘Te Peril.’ ” Federal Reser ve Bank o St. Louis Review, September/October 2010, Vol. 92,No. 5, pp. 339-52.
2 Also, see Hursey, im; and Wolman, Alexander L.“Monetary Policy and Global Equilibria in a Produc-tion Economy.” Federal Reserve Bank o RichmondEconomic Quarterly, Fourth Quarter 2010, Vol. 96,No. 4, pp. 317-37.
3 See Bullard, James. “QE2 in Five Easy Pieces.” Speechat the High Pro le Speaker Series, New York Society o Security Analysts, New York City, Nov. 8, 2010.
4 Research at the St. Louis Fed suggests that quantita-tive easing programs in the U.S. can have interna-tional e ects (e.g., a reduction in long-term oreignbond yields), as well. See Neely, Christopher J. “TeLarge-Scale Asset Purchases Had Large Interna-tional E ects.” Federal Reserve Bank o St. LouisWorking Paper 2010-018C, January 2011.
easing o monetary policy in that the imprinton the nancial markets looked just like astandard, aggressive monetary policy easing.Furthermore, the disin ationary trend o 2010 has apparently been reversed, and theU.S. economy seems to have avoided theJapanese-style outcome. Although a rule-like approach would have been pre erable
rom my point o view, rather than indepen-dent, isolated decisions with large amountso purchases, the impact o quantitativeeasing on macroeconomic and nancialconditions showed that the Fed has plenty o ammunition to carry out stabilizationpolicy even when the policy rate cannot belowered urther.
Te heavy reliance o businesses on com-modities is illustrated by the story o JohnAnton, ounder and owner o Anton Sport,a wholesaler o athletic apparel in empe,Ariz. Anton, who normally keeps on hand30 boxes o cotton -shirts as inventory,was reported this February by Te Wall Street Journal to be sitting on 2,500 boxeso cotton -shirts, unded via a $300,000loan.2 Te impetus? A 90 percent increasein the price o cotton over 2010.
Currently, commodity prices are makingheadlines as much or the size o the priceincreases as or the simultaneity o pricehikes across all types o commodities.Figure 1 reveals that, prior to the globalrecession, upward price trends took hold ina variety o commodities. Te nancial cri-sis and ensuing recession induced an acutedecline rom the 2008 peak in prices. Butbeginning in 2009, the prices o all typeso commodities began to rise once again atastronomical rates.
Tis synchronization o price movementsacross a range o commodities has ostered,in part, the assertion that the commodity price boom is a bubble, driven primarily by near-zero interest rates and excessivespeculation in commodity utures markets.Te counter argument is that market un-damentals—supply and demand orthe commodities themselves—can ully explain the price gains. Ultimately, under-standing the sources o the price gains is
essential or determining the proper policy response, i any.
Arg ments or Mar et F ndamenta s
In the absence o “irrational exuberance,”the price o any good or asset should bedriven by supply and demand. On both thesupply and demand side o commodities,there is no shortage o shocks to explain, atleast in part, recent price gains.
Negative Supply ShocksFor crops and many other commodities,
annual production is largely at the discretiono Mother Nature. With respect to agricul-tural commodities, a combination o badbreaks rom Mother Nature and stock-to-use ratios at already historic lows seems toexplain much o the price increases.
Pre-existing stocks are a key source o stability in commodity markets. Whenstocks are low relative to use, the market isless able to absorb pressures rom supply
disruptions or unexpected demand; theresulting pressure on prices is much stron-ger. A survey o commodities characterizedby rising prices uncovers many stock-to-useratios at historic lows.
In a report on the pre-recession spikein ood prices, the Food and AgricultureOrganization o the United Nations (FAO)identi ed numerous reasons why stock levels have been alling by an average rateo 3.4 percent per year since the mid-1990s.3
Reasons included declines in the reservesheld by public institutions, developmento other less costly instruments o risk management, increases in the number o countries able to export, and improve-ments in in ormation and transportationtechnologies. Further, the FAO oundstrong evidence that lower stock levels atthe beginning o the marketing season wereassociated with higher prices throughoutthe season, implying initial conditions in
“tight” markets matter. Compounding thise ect is urther empirical evidence that theprice impact o low stocks becomes magni-
ed when stocks reach critically low levels.For all o these reasons, low stocks in
ood and other crops mean that the weatherdisruptions aced in 2010 were all thatmuch more signi cant. For example, the47 percent increase in wheat prices in 2010was largely attributable to drought in Russiaand China and to oods in Canada andAustralia. High cotton prices can be traced,
in part, to oods in China (the largestproducer) and Pakistan (the ourth-largestproducer).
In many cases, the high prices in onemarket have spilled into other marketsbecause o the competition between crops
or the same land and growing resources.Farmers are choosing to grow the cropsthat are in shortest supply with the highestprices, o en introducing shortages in otherdisplaced crops.
With respect to nonagricultural commodi-ties, the challenge o suppliers is less a resulto temporary negative shocks than it is aresult o rapidly expanding global demand.
Growing Demand
Te convergence in income between devel-oping and advanced countries represents asigni cant driver o demand growth or com-modities: Representative o the trend, morethan 90 percent o the increased demand oragricultural commodities over recent yearshas originated in developing countries.
For commodities such as metals, thisadditional demand can take time to ully accommodate. Figure 2 reveals the incred-ible pace at which demand or metals likealuminum and copper has grown in the two
most populous emerging countries: Chinaand India. Tis huge demand growth isa major contributor to the InternationalCopper Study Group’s ndings that world-wide demand or re ned copper exceededworldwide supply by 480,000 tons over the
rst nine months o 2010.4 Te mismatchbetween supply and demand has unsur-prisingly taken a large toll on inventories,cutting them by more than hal , rom 1.1million tons in 2001 to 412,000 tons by September 2010.
Continued strong growth in emergingcountries, complemented by economicrecovery in the United States, Japan andEurope, is expected to continue to putupward pressure on prices o metals.According to Bloomberg News, 13 o 14industry analysts who were surveyedexpected a copper shortage this year.
While exploration and investment inmining operations are under way, muchtime and money will be required be ore
new mines are operational. In the wordso U.S. Geological Survey specialist DanielEdelstein, “Mines aren’t just like actories,where you just ip a switch.”
With respect to agricultural markets, theFAO is correct to point out that increaseddemand due to population and incomegrowth is largely predictable. Bio uels,however, are cited as a new and persistentshock to ood demand.5 Figure 3 reveals anunmistakable recent shi in the relationshipbetween oil prices and the price o popularbio uel crops, such as corn ( or ethanol) andsoy ( or biodiesel). Te enormous size o energy markets compared with agriculturalmarkets means that energy-related demandis capable o absorbing near-limitlessamounts o surplus crops, e ectively placinga oor below ood prices. While great or
armers, this is unwelcome news or theimpoverished and malnourished popula-tions o the world. Te e ect o bio uelsis also not limited to crops used in theirproduction. Bio uel production representsan alternative use o land, which a ects allagricultural products.
Te outlook in oil markets, which drivesdemand or bio uels, is not particularly promising either. According to a recentreport rom the International Monetary Fund, oil demand in emerging markets isquickly catching up to demand in advanced
countries a er years o signi cantly lowerconsumption rates by the ormer. 6 Com-pounding this situation, production con-straints in current exporting countries arestarting to bind, as oil elds have reachedmaturity. One source o relie may come inthe orm o shale oil, in which the UnitedStates is rich. But extraction rom shale willnot become sustainable until the price o oilpromises to stay above $80-105 a barrel. 7
Overall, there is no doubt that undamen-tal shocks to supply and demand in com-
modities, both transitory and persistent,can account or signi cant price pressuresin these markets. Some, however, remainunconvinced that these undamental shocksare enough to explain the entirety o priceincreases. Instead, they place some blameon a bubble in commodity prices.
Arg ments or a B bb e
An asset bubble is characterized by pricesdetached rom undamentals, instead driven
by the anticipation o pro ting romhigher prices tomorrow.
Commodity markets, however, do notmeet the usual theoretical criteria ora bubble. Arguments or a speculativebubble ocus primarily on one market-place or commodities: the uturesmarket. Commodity utures markets arewhere both commercial and noncommer-cial traders can buy and sell standard-ized contracts or delivery o a speci edquantity o goods at a speci ed date in the
uture. Tese contracts are short-terminstruments that have ew constraints onshort-selling (betting on price decreases)and that are easy to arbitrage (pro t risk-
ree rom mispricing). In contrast, theory holds that bubbles are limited to marketssuch as real estate, where the good inquestion has a long li espan, is hard to sellbe ore you own, and buying and selling iscostly in terms o time and money.
Still, some believe that a bubble isorming in commodities due to either
expansionary U.S. monetary policy and/or record ows o investment unds intocommodity utures. Tese possibilitieswarrant care ul consideration.
The Role of Expansionary U.S. Monetary Policy
Te primary means by which expan-
sionary monetary policy in uences com-modity prices is by decreasing the cost o holding inventories. Anton, the apparelwholesaler, provides a good example.One component o the cost o holdinginventory is the prevailing interest rate.Expanding inventory means borrow-ing money, as in the case o Anton, orsacri cing the return that one couldearn rom investing the money. Near-zero interest rates, as currently exist inthe United States, signi cantly decrease
the cost o holding inventory and, thus,increase demand or commodities. Inthis context, inventory buildups, suchas Anton’s, can be interpreted as symp-tomatic o overly loose monetary policy.Broad declines in aggregate commodity inventories, however, cast doubt on thecurrent importance o this e ect.
Te quotation o international com-modity prices in dollars opens a sec-ond means or U.S. monetary policy in
particular to in uence commodity prices.When the dollar depreciates, goods pricedin dollars become more a ordable to oreignconsumers, all else equal leading them toincrease consumption and bid up the priceson these goods. Tis argument is countered,however, by the observation that commod-ity prices rose signi cantly over recent yearsregardless o the currency quoted in.
Te rather recent argument that has beenput orth is that historically low U.S. inter-est rates have increased commodity pricesby driving investment unds into othermarkets, including the nancial markets o emerging countries, to seek higher returns.Te evidence, however, is ounded mostly
on correlation and largely lacks a credibletransmission mechanism. Completing thetheory o how an in ow o capital to emerg-ing markets in ates commodity pricesrequires a link between the in ow o oreigninvestment and a broad expansion in emerg-ing market credit. Ultimately, the bankingsystems o the developing countries receiv-ing the in uxes o capital must transmit the
unds into the general economy. But the
skepticism that developing countries likeBrazil, Tailand and Indonesia have showntoward much o the capital in ows, labelingthe unds as “hot money” seeking short-term returns, places uncertainty over theextent that capital in ows are unding bid-ups in commodity prices among developingcountries.
Te impact o increased speculationin commodity utures markets, perhapsexacerbated by low traditional investmentreturns, has been an area o intense research
in recent years, however.
The Potential Costsof Excessive Speculation
Just as well-documented as the large gainsin commodity prices prior to the recessionis the contemporaneous large in ux o capital into the commodity markets, namely in long-only index unds.8 According toBarclay’s, index und investment in com-modities increased rom $90 billion in early
2006 to just under $200 billion by the endo 2007. Te proposed link between large
ows o capital into commodity marketsand increases in current prices appeals tocommon sense: Speculative demand orcommodity-based assets increases demand
or the underlying commodity, increasingits price. A second practically oundedrationale or why excessive speculationmust have played a role in rising commodity prices is embodied by a U.S. Senate com-mittee sta report in 2006: “Te traditional
orces o supply and demand cannot ully account or [energy price] increases.” 9
Despite these straight orward proposi-tions, however, the true impact o specu-
lative in ows on underlying commodity prices remains debatable. A technicalreport prepared or the Organisation orEconomic Co-operation and Development(OECD) o ers a use ul examination o theresearch done on both sides. 10 In particular,the authors pointed out both logical and ac-tual inconsistencies within the argument ora speculation-induced bubble in commod-ity prices. Logical inconsistencies include
a tenuous link between speculative in owsand demand or the underlying commodity and doubt over the extent that index undinvestors could arti cially increase uturesand cash prices while only participating inthe utures market and not the spot market,where commodities are sold or immediatedelivery. Factual inconsistencies are numer-ous. For example, inventories should haverisen between 2006 and 2008 accordingto the bubble theory, but they actually ell.Other reasons or discounting this theory
include:• arbitraging index-fund buying is fairly
easy due to its predictable nature,• commodity prices rose in markets with
and without index unds,• speculation was not excessive a er
accounting or hedging demand, and• price impacts across markets were not
consistent or the same level o indexund activity.
In addition to their own analysis, the
authors o the OECD report reviewed ourstudies supporting a pre-recession com-modity bubble and ve studies discountinga bubble. Te authors concluded that “theweight o the evidence at this point in timeclearly tilts in avor o the argument thatindex unds did not cause a bubble in com-modity utures prices.” O the studiessupporting a bubble, they write, “Tesestudies are subject to a number o importantcriticisms that limit the degree o con denceone can place in their results.” Still, theOECD report contains an important caveatregarding the markets most o en linked to aspeculative bubble: “Te evidence is weakerin the two energy markets studied becauseo considerable uncertainty about the degreeto which the available data actually re ectindex trader positions in these markets.”
Sorting out the bubble arguments hasextremely important policy implicationsgoing orward.
Are Po ic Responses Req iredin Commodit Mar ets?
Te most important thing to rememberwith respect to commodity markets is thatthey are volatile. Te traditional decisiono central banks to ocus on core in ation,which excludes ood and energy, is easy tounderstand in the context o recent move-ment in rubber markets.
During 2010, the price o rubberincreased by 114 percent. Te run-up inthe price was largely attributed to badweather, low stocks and growing demand
rom China’s automobile industry. Aroundthe end o 2010, many investors remainedbullish on rubber prices due to expecta-tions o continuing strong demand. Indeed,the real price o rubber reached a historicpeak in the middle o this February. Yetonly a month removed rom that peak, theprice ell more than 30 percent in a mat-
ter o weeks, and the Tai government wasdiscussing price supports or rubber. Teprice drop was due to uncertainty overglobal demand, stemming rst rom unrestin the Middle East and, subsequently, theearthquake and resulting tsunami in Japanand their uncertain e ects on the demand
or rubber tires rom Japanese carmakerslike oyota, Honda and Nissan. Tis dropwas then ollowed by a 23 percent increasein the price over the second hal o March
as Tailand, the largest producer o rubber,ultimately intervened to buy up domesticrubber supplies and support prices, whilesimultaneously telling armers to restrictsupplies in an e ort to bid prices back up.
Not only are large movements in commod-ity prices common, but they are o en linkedto inherently unpredictable events. Just inthe past ew months, cotton prices ell by 25percent and oil had its largest one-day dropin two years. o try to design policy aroundcommodity prices would require abruptabout- aces and would detract rom a centralbank’s goal o bringing stability to markets.
More pertinent questions with respect tocommodity markets are:• Is strong regulation in futures markets
needed?• Are large subsidies on biofuels good policy?• Should U.S. monetary policy take into
consideration global economic conditions?Some countries, like India, have already
begun to regulate commodity utures mar-kets; other countries, including the UnitedStates, have debated the issue. Both thosewho believe in a speculative commodity bubble and those who do not can agree thatproperly unctioning commodity uturesmarkets are integral to the real economy because they allow those who do not wish tohold the risk o uture price movements tosell that risk to will ing parties. Te OECD
report provides a reminder that index undinvestors are an important source o liquid-ity and o risk absorption or these markets.Pushing such investors out o the marketcould result in huge costs, which must beweighed against the evidence that theiractivity is hindering, and not enhancing,the proper unctioning o these markets.
With respect to bio uels, potentialnegative e ects, such as reversing a 30-yeardownward trend in real ood prices, are o particular relevance because these markets
are currently highly dependent on govern-ment subsidies. Brazil’s ethanol rom sugarcane is the only bio uel whose production is viable without government subsidies. In theUnited States, subsidies on ethanol increasethe price that processors can a ord to pay
or corn and break even (a unction o oilprices) by $63 per ton. Tis compares withan average price o corn in 2005 (predatingheavy investments in bio uel) o $75 per tonand a price o $163 per ton that processors
can already a ord to pay and break evengiven crude oil prices o $100 per barrel.
Government support o the industry ismotivated by bene ts, such as energy inde-pendence and a reduction in the environ-mental impact, that accrue to society butcannot be internalized by processors. Butrecent li e-cycle analysis o bio uels—ananalysis that takes into account the extraland needed to grow crops and the produc-tion process—raises questions about theenvironmental bene ts. Te question iswhether there may be less-costly and more-e cient ways to achieve the same policy goals. Te long-run success o bio uels islikely to hinge on the development o sec-ond-generation uels, which can make useo more parts o the crop, as well as bio uelsbased on highly e cient algae.
Te nal question regarding the consider-ation o global economic conditions in U.S.monetary policy debate will require muchmore convincing evidence be ore a rmconclusion can be reached. I expansionary U.S. monetary policy is transmitted glob-ally to economies in danger o overheating,which in turn bids up commodity pricesand, hence, increases price levels back athome, then U.S. monetary policy shouldcare about output gaps around the world.At the same time, the mere correlation o commodity price increases with loose U.S.
monetary policy, without any convincingempirical evidence or theoretical mecha-nisms or this avenue, is not enough todetermine that U.S. policy decisions should
actor in economic conditions rom LatinAmerica to Europe, rom Asia to A rica.
Ultimately, the greatest lesson rom recenttrends in commodity prices may be thereminder that economics is ounded on theassumption o a world with unlimited wantsand limited resources. A world with a grow-ing population and ever-increasing income
parity implies a world with ever-increasingcompetition or resources.
Luciana Juvenal is an economist and Brett Fawley is a senior research associate, bothat the Federal Reserve Bank o St. Louis. Seehttp://research.stlouis ed.org/econ/juvenal/ or more on Juvenal’s work.
E N D N O E S
1 See O’Donnell.2 See Pleven and Wirz.3 See Food and Agriculture Organization o the
United Nations (2009).4 See Davis.5 See Food and Agriculture Organization o the
United Nations (2008).6 See International Monetary Fund (2011).7 See Engemann and Owyang.8 “Long-only” re ers to t he act that t hese index
unds make only buy and sell decisions and donot short utures contracts.
9 See Senate Report 109-65. 10 See Irwin and Sanders.
Engemann, Kristie M.; and Owyang, Michael .“Unconventional Oil Production: Stuck in aRock and a Hard Place.” Te Federal ReserveBank o St. Louis’Te Regional Economist ,July 2010, Vol. 18, No. 3, pp. 14-15.
Food and Agriculture Organization o the UnitedNations. Te State o Food and Agriculture2008. Rome: Food and Agriculture Organi-zation o the United Nations, 2008.
Food and Agriculture Organization o theUnited Nations. “Te State o Agricultura lCommodity Markets: High Food Prices andthe Food Crisis—Experiences and L essonsLearned.” 2009.
Internationa l Monetary Fund. “World EconomicOutlook, April 2011. ensions rom the
wo-Speed Recovery: Unemployment,Commodities, and Capital Flows.”
Internationa l Monetary Fund. “RegionalEconomic Outlook, October 2010. As ia andPaci c: Consolidating the Recovery andBuilding Sustainable Growth.”
Irwin, Scott H.; and Sanders, Dwight R. “TeImpact o Index and Swap Funds in Commod-ity Markets.” A technical report prepared orthe Organisation or Economic Co-operationand Development, June 2010.
O’Donnell, Jayne. “Wal-Mart CEO Bill SimonExpects In ation.” USA oday, April 1, 2011.
Pleven, Liam; and Wirz, Matt. “Companies StockUp as Commodities Prices Rise.” Te Wall Street Journal , Feb. 3, 2011.
Senate Report 109-65, “Te Role o MarketSpeculation in Rising Oil and Gas Prices:A Need to Put the Cop Back on the Beat.”Sta report prepared by t he PermanentSubcommittee on Investigations o theCommittee on Homeland Security andGovernmental A airs, U.S. Senate,June 27, 2006.
he 2007-09 recession had a severe impacton the U.S. labor market. During the
recession, more than 89 million employeeslost their jobs, while ewer than 82 millionwere hired.1 Te unemployment rate spikedto a 27-year high o 10.1 percent in October
2009. Since then, the labor market has expe-rienced a slow recovery; the unemploymentrate still stood at 9.1 percent in May.
In the 2010 annual report o the FederalReserve Bank o St. Louis, David Andol attoand Marcela Williams suggested that search“ rictions” might explain why the unemploy-ment rate remained high even while job
Te Mismatch betweenJob Openings and Job Seekers
By Maria E. Canon and Mingyu Chen
m b b k ll l q ll v j b k ll
l k .openings appeared to have increased duringthe recent recovery. One type o rictionthat they mentioned relates to employer-employee pairings: Each job and worker hasidiosyncratic characteristics that make some job-worker pairings more productive thanothers. As employers and workers usually cannot anticipate where the best pairing islocated, they must expend time and resourcesto search out the best matches.
Mismatch can be interpreted as a poormatch between the skills and locationrequired to ll vacant jobs and the skillsand geographic pre erences o unemployedworkers. Te idea, also known as structuralimbalance, was rst identi ed by a groupo European economists in the 1970s, whenthey were struggling to understand theconsistently high unemployment rate insome European countries. 2
In general, skills can be represented in
di erent contexts, such as industries, occupa-tions and educational levels. Geographiccharacteristics can be measured at di erentlevels, such as metropolitan statistical areas(MSAs), states and, at an even larger level,census regions. Economists have recently
paid close attention to mismatch and haveinvestigated whether it is causing the cur-rently high unemployment rate in the U.S.
Some evidence suggests that mismatchmight have increased since the recessionstarted. Te gure shows the averagemonthly share o vacant jobs and share o employment lost by industry rom December
2007 to February 2011.3 Most new positionshave been created in some sectors, while most job loss has been concentrated in others.Since these new jobs usually require di erentskills than what unemployed workers romdi erent sectors have, rms and unemployedworkers may take longer to nd their bestmatches. For example, over 50 percent o the jobs lost between December 2007 andFebruary 2011 were in manu acturing and
construction, while more than 90 percento new positions opened in other industries.Te education and health sector has experi-enced steady employment growth since therecession started; 20 percent o all job open-ings have occurred in this sector.
In the rest o the article, we review therole o two types o mismatch (skill andgeographic) in explaining the increase inunemployment that occurred during anda er the 2007-09 recession.
S i Mismatch
Economists Ayşegül Şahin, Joseph Song,Giorgio opa and Giovanni Violante recently derived mismatch indexes rom an economicmodel. 4 In their ramework, the aggregatelabor market is comprised o many smalllabor markets, categorized by skill levelsor working locations (e.g., industries andMSAs). Şahin and others de ne mismatch asthe distance between the observed allocationo unemployed workers across sectors andthe “optimal” allocation. Te optimal alloca-tion o unemployed workers is the allocationthat, given the distribution o vacancies inthe economy, would occur i there were reemovement o workers across labor markets.Te authors’ indexes allow them to quanti y not only the level o mismatch but also the
proportion o the increase in unemploymentthat can be attributed to mismatch.
Using ve industries as divisions o theaggregate labor market, Şahin and herco-authors ound that the raction o unem-ployed workers misallocated increased by 10 percentage points during the 2007-09recession; the raction then dropped butremained at a level higher than its prereces-sion level. But this increase in mismatch canexplain only between 0.4 and 0.7 percentagepoints o the total increase o ve percent-
age points in the unemployment rate romthe beginning o 2007 to the middle o 2009.Tere ore, although skill mismatch increasedduring the recession and in uenced unemploy-ment to some degree, it is not the main sourceo the increase in the unemployment rate.
Geographic Mismatch
Te 2007-09 recession was accompaniedby a steep decline in housing prices. Someeconomists and commentators have argued
Te Foreclosure Crisis in 2008:Predatory Lendingor Household Overreaching?
W atching southern Florida home pricesspiral out o reach, Mr. Briar decided
to take the plunge in 2004 and buy his rsthome. Te mortgage broker he workedwith encouraged him to enter into a 2/28contract, in which the interest rate is xed
or the rst two years and then resets to ahigher oating rate. Mr. Briar bought thehome, and the mortgage broker trans erredthe loan to Wall Street, where it was pack-aged and securitized into a collateralizeddebt obligation (CDO). Mr. Briar struggledto pay his mortgage even during the rst
two years. Meanwhile, Florida home pricesplunged, and, eventually, Mr. Briar perma-nently de aulted on his loan. Te servicingbank oreclosed nine months later.
Although Mr. Briar is a ctitious person,this story has played out or millions o households over the past ew years. Did Mr.Briar overreach by taking on too much hous-ing debt, or was he duped by Wall Street?Te answer is di cult to ascertain because
it ultimately depends on the intentions o the borrower and the lender. A er the act,a lender would hardly admit to deceiving aborrower, and the borrower would be morethan willing to place at least some o theblame or the oreclosure on the lender.
Certainly, both predatory lending andhousehold overreaching occurred during thesubprime housing bubble. But it is importantto identi y the primary reason or the oreclo-sure crisis because the policy implications are
vastly di erent. I predatory lending was theprimary culprit, strong consumer protectionlaws like those in the Dodd-Frank law mightbe su cient to avoid a uture oreclosurecrisis; that’s because such laws would preventWall Street banks rom making high-risk loans that borrowers could not possibly a ord. I household overreaching was theprimary culprit, preventing another oreclo-sure crisis is a much more complex policy challenge. A return to high appreciation inhome prices could again set o dynamicsin which even borrowers with decent credit
would overreach and end up in homes they ultimately couldn’t a ord. Te only com-prehensive solution might be to prevent the
ormation o asset price bubbles, a solutionthat would require policymakers, such as thecentral bank, to recognize and de ate suchbubbles when they occur.
o distinguish between the predatory lending and overreaching hypotheses, wetapped two nationwide data sources to
analyze the characteristics o households inoreclosure. Because private motivationswere unobservable, we argue that householdswith low income and education levels shouldbe the most vulnerable to predatory lend-ing practices because such borrowers, allelse equal, are more likely to have a poorerunderstanding o the contract terms at thetime o origination. In contrast, householdsmost susceptible to overreaching are thosethat have high economic aspirations relative
to their current income and net worth; thesehouseholds could already have relatively highincomes and be well-educated.
Pro es o Forec osed Ho seho ds
Te data used in our analysis o oreclosedhouseholds came rom two sources. Realty-
rac compiles nationwide data on homesin oreclosure. Acxiom compiles data onmillions o U.S. households each quarter andsegments households based on economic,demographic and consumption patterns. oobtain a pro le o oreclosed households, wecombined these two large datasets by house-hold or the third quarter o 2008. Te data-set contains more than 40 million recordsand more than 200,000 oreclosures.
Figure 1 presents key statistics rom our
dataset on households in oreclosure along-side households not in oreclosure. De aultedhomes were more expensive, on average. Temedian market value o homes in oreclosurewas $242,400 versus $199,129 or homes notin oreclosure. As expected, the medianloan-to-value ratio was much higher onde aulted properties, at 96 percent, which wasmore than 30 percentage points higher thanon nonde aulted properties. Homes in ore-closure also were slightly newer and smallerin terms o square ootage.
Household characteristics, shown in thebottom panel, reveal that households inoreclosure had slightly ewer members and
were signi cantly younger. Te medianhead-o -household age or a oreclosedhousehold was 44, eight years younger thanthe median or households not in oreclosure.Heads o households in oreclosed propertieswere less likely to be married and more likely to be single. Tey had lower incomes andmuch shorter length o residence. Although
mean years o education were similar at justover 14, households in oreclosure had amedian 12 years o education compared witha median o 16 years or households not in
oreclosure.Because we were interested in identi ying
the characteristics o households that wereresponsible or a disproportionate numbero oreclosures, we looked beyond the simpleaverages described above. PersonicX Li eStage Segmentation is an Acxiom classi ca-tion scheme that divides households into21 li e stages based on marital status, numbero children in the household, employmentstatus and other socio-economic characteris-tics.1 A number and letter correspond to thename o each group listed in Figure 2. Tenumber corresponds to the age o the group,with lower numbers representing youngerdemographics; the letter approximates thegroup’s cultural generation. Groups endingin B represent the Baby Boomers, while Xand Y represent Generation X and Genera-tion Y. M represents the Mature generation,mostly those in their 50s and 60s, and S rep-resents Seniors, most o whom are retired.
o see which o the 21 PersonicX groupscontributed the most disproportionately tothe oreclosure crisis, we calculated the shareo total oreclosures represented by eachgroup and the share o all households rep-resented by each group. We subtracted the
household share rom the oreclosure share toderive the “excess oreclosure shares” o eachgroup. Group 07X, or example, accounted
or 5.52 percent o all households but 11.3percent o all oreclosures. Te excess shareo oreclosures is the di erence o these tworatios, or 5.78 percentage points. Figure 2plots the 11 PersonicX Groups with the high-est excess oreclosure shares.
Figure 2 shows that excess oreclosurescame primarily rom younger, relatively afu-ent households, a nding more consistent
with the overreaching hypothesis. In parti-cular, the group with the largest number o excess oreclosures was 07X,Cash & Careers.Tis Generation X group was the mostprosperous o the generation o adults bornin the mid-1960s and early 1970s. Out o the
rst 10 PersonicX groups with excess ore-closures, Cash & Careers members ranked
rst in average household income ($59,500),net worth and years o education (14.8). Tesecond most-overrepresented group in terms
Not in Foreclosure In Foreclosure
Mean Median Mean Median
Property Characteristics
h m k v $278,115 $199,129 $290,653 $242,4
h p c a $198,598 $140,000 $253,650 $199,9
l v 64.6% 65.0% 90.7% 96.0%
y h b 1969 1974 1972 1978
h s z ( q ) 2,376 1,907 1,554 1,52
Household Characteristics
h d s z 3.3 3.0 2.9 2.0
a i c $55,700 $51,500 $51,241 $48,800
y ed c 14.8 16.0 14.1 12.0
a 53.1 52.0 45.1 44.0
l r d c 9.1 9.0 5.3 4.0
n C d 1.4 1.0 1.5 1.0
m d 70.7% 56.2%
s 25.7% 36.9%
u.S. Propert and Ho seho d Characteristics b Forec os re Stat s
FIGURE 1
sourCes: acx , r t c d ’ w c c .
o excess oreclosures was 02Y, aking Hold .Tese were Generation Y households with anaverage age o 27.8 years, second-highest aver-
age income ($55,500), third-highest net worthand h-highest education level (14.1 years).Tese two groups’ characteristics were consis-tent with our expectations o households thatare most likely to overreach.
Te two groups in Figure 2 that were mostlikely to be victims o predatory lending wereGroup 01Y, Beginningsand Group 06X, Mixed Singlesbecause these groups ranked ninthor 10th in income, net worth and education.Yet these groups ranked seventh and eighth,
Excess Forec os re Percentages b PersonicX Gro p or u.S. Ho seho ds
FIGURE 2
07X Cash & Careers
02Y Taking Hold
03X Transition Blues
09B Boomer Singles
05X Gen X Parents
08X Jumbo Families01Y Beginnings
06X Mixed Singles
10B Mixed Boomers
04X Gen X Singles
14B Our Turn
5.78 percentage points
3.66
2.97
2.40
1.78
1.701.43
1.24
1.19
1.18
0.05
sourCes: acx , r t c d ’ w c c .
t w w xc c . t c
c acx ’ p cX l ss . i ,
w . t d : b=b
b , X=g X d y=g y.F x , 07XC sh & C reers cc d 5.52 c d 11.3 c
and jointly, they accounted or just 2.67percentage points o excess oreclosuresrelative to 9.44 percentage points or groups07X and 02Y.
Rather than rely solely on Acxiom’sgroupings, we also separated all the house-holds into quadrants based on income andeducation to identi y the most leveragedhouseholds in each quadrant based on theirloan-to-income ratio. We conjectured thatthe most over-leveraged households inthe low-income, low-education (bottom)
quadrant were more likely to be victims o predatory lending, while the most over-leveraged households in the high-income,high-education (top) quadrant were morelikely to have overreached. Our tests showedthat the most-leveraged households in thetop quadrant were statistically more likely toenter oreclosure than the other householdsin the same quadrant. Tis pattern was nottrue, however, or households in the bot-tom quadrant. Once again, overreaching
appeared to be the more important explana-
tion o mortgage oreclosure.
Geographic Patterns o Forec os res
In addition to household pro les, ourhypotheses also have di ering implications
or the geographic distribution o oreclo-sures. Te predatory lending hypothesispredicts that the geographic distribution o
oreclosures will re ect the spatial distribu-tion o low-income and low-educated house-holds because bankers (or their brokers) willseek out households most easily deceived,
regardless o the household’s location. Incontrast, the overreaching hypothesispredicts that bubble dynamics will be theimportant actor explaining the oreclosures.Tis hypothesis implies that oreclosurerates will spike in speci c “hot spots” wherehouseholds and speculators bid up prices inan e ort to buy more-expensive homes be orethese homes become una ordable.
We identi ed real estate hot spots usingdata rom the Federal Housing Finance
Ann a ized Forec os re Rates, 2008:Q3
sourCe: r t c
FIGURE 3
t c c c w c c d k
z d d
w c d dq 2008 d d d d .
Maine
New York
Pennsylvania
Massachusetts N e w
H a m p s h i r e
V e r m
o n t
Rhode IslandConnecticut
New Jersey
DelawareMaryland
WestVirginia
Virginia
NorthCarolina
SouthCarolina
Georgia
Florida
OhioIndianaIllinois
Kentucky
Tennessee
Washington, D.C.
Alabama
Mississippi
Arkansas
Louisiana
Texas
Oklahoma
MissouriKansas
NebraskaIowa
Minnesota
South Dakota
North DakotaMontana
Washington
Oregon Idaho
Nevada
California
Utah
Arizona
Alaska
New Mexico
Wyoming
Colorado
Wisconsin
Michigan
Hawaii
0.00 to 1.37 No data available9.63 to 11.00 8.25 to 9.62 6.88 to 8.24 5.50 to 6.87 4.13 to 5.49 2.75 to 4.12 1.38 to 2.74
1 A list o the 21 PersonicX li e stages and theirdescriptions is available rom the Acxiom website at www.acxiom.com/products_and_services/Consumer%20Insight%20Products/segmentation/Pages/index.html
R E F E R E N C E S
Minsky, Hyman P. Stabilizing An UnstableEconomy. New York: McGraw Hill, 2008.Originally published by Yale University Press,New Haven, Conn., 1986.
Agency House Price Index between 2000 and2007. Te areas with the most signi canthome appreciation are Florida and the statesin the Southwest and in the Northeast.
Figure 3 is a map o oreclosure ratesby state or the third quarter o 2008. Teoverreaching hypothesis suggests that thereshould be a strong correlation between thestates with the greatest price increases andthe states with the highest oreclosure rates.Indeed, the concentration o oreclosures inthe Southwest and in Florida is consistentwith overreaching as a more importantexplanation than predatory lending orthe oreclosure crisis. Te main outliers inFigure 3 are the Great Lakes states, such asMichigan, Illinois, Indiana and Ohio, all o which experienced moderate home-priceappreciation but relatively high oreclosurerates. Foreclosures in these states are more
likely driven by a weak economy rather thanby housing price bubbles.
o more rmly support this visual evi-dence, we ranked all o the 50 states by homeprice appreciation (between 2000 and 2007)and oreclosure rates (in 2008) to evaluatetheir statistical correlation. Te overreachinghypothesis suggests that these two char-acteristics should be positively correlated.Indeed, or all the states, the correlation is0.23—positive as the overreaching hypothesissuggests, though not statistically di erent
rom zero. When we exclude the Great Lakesstates, however, the rank correlation rises to0.43 and is statistically signi cant. Again,the evidence is more consistent with the over-reaching hypothesis than with the predatory lending hypothesis.
Po ic Response to Asset B bb es
By combining household oreclosure datarom Realty rac with household data rom
Acxiom, we were able to create a pro le o households in oreclosure during the early
stages o the nancial crisis. We ound thatmany oreclosed households were young withrelatively high income and education levels.Moreover, geographic oreclosure patternswere consistent with bubble dynamics asillustrated by the positive correlation betweenhome-price appreciation and subsequent
oreclosure rates. Te weight o the evidencesupports the overreaching hypothesis.Consequently, strong predatory lendingrestrictions, while desirable, would likely
be insu cient to avoid a uture oreclosurecrisis should another housing bubble emerge.
In our view, the ultimate underlying causeo the oreclosure crisis was the emergence o a signi cant housing price bubble and its sub-sequent collapse. Un ortunately, preventingasset price bubbles is a much more complexpolicy problem to address than protectingconsumers rom predatory lending.
Te late economist Hyman Minsky arguedthat capitalist economies go through lever-age cycles, in which credit access becomesprogressively easier as an economy growsstrongly. Te success o lenders and rms inthe good years, combined with appreciatingcapital assets, reduces the perception o risk and encourages increasingly riskier nanc-ing. Financial innovation exacerbates theleverage cycle as nancial rms devise newways to extend credit. Eventually, asset prices
peak and then begin to decline, nancialinstability emerges and latent systemic risk isunleashed in a nancial crisis.
Tis leverage cycle, which Minsky calledthe nancial instability hypothesis, may beinherent to the capitalist system. Minsky’sthesis might portray the subprime nancialcrisis quite well, but it also would suggestthat uture crises can result rom assetbubbles in other sectors o the economy,not just housing.
I capitalist economies are subject to peri-
odic asset price bubbles, Minsky suggestedthat policymakers take steps to eliminatebubbles that threaten to become systemically important. Tis, o course, requires the abil-ity to 1) recognize an asset bubble, 2) classi y the bubble as a systemic risk to the economy and 3) curb the ormation o the bubble eitherthrough monetary policy actions or throughmore-targeted interventions, such as higherbank capital requirements or more stringentmortgage underwriting criteria.
William R. Emmons is an economist at theFederal Reserve Bank o St. Louis. Seestlouis ed.org/emmonsvitae or more on hiswork. Kathy Fogel, Wayne Y. Lee, Liping Ma,Deena Rorie and imothy J. Yeager are at theSam M. Walton College o Business at the Uni-versity o Arkansas. See http://waltoncollege.uark.edu/fnn/PredatoryLendingOverreaching. pd or the complete research paper.
Central to economic events o recenttimes were the rapid increases in house
prices a er 1995 and the ensuing downturnin those prices around 2006-07. Naturally,the importance o accurate measurement o house price trends can hardly be overem-phasized. Several prominent house priceindexes have been developed or the UnitedStates. Tese include the National Associa-tion o Realtors (NAR) median index, theCensus Bureau median index, the S&P/Case-Shiller national index, the CoreLogicindex and the Federal Housing FinanceAgency (FHFA) index. 1 Each di ers in
methodology, in its emphasis on the varioussegments o the housing market or both. othe casual observer, the di erence in pricechanges recorded on each o the indexescan be perplexing. Tere ore, knowing howthe indexes di er rom one another can beinstructive as to which index to ollow.
Housing price indexes are calculated by tracking home prices in a given region over aperiod o time. Ideally, one would track the
price o a random sample o houses. How-ever, this method has operational problemsbecause, at any particular point in time, notall houses are or sale; additionally, there may be variations in the type o houses sold. I one merely tracked the price o homes soldover time (e.g., as is ound in median houseprice indexes, such as the NAR and theCensus indexes), observed changes could bedue to changes in the composition o homessold as opposed to changes due to market
conditions. Dealing with houses that di erin “hedonic” characteristics—such as thesquare ootage, number o bedrooms anddistance rom city center—can be tricky.
o deal with these issues, economistshave adopted a “repeat sales” methodology,which measures price changes o the samehouse between a previous and current sale. 2 Examples o repeat sales indexes include theCase-Shiller, CoreLogic and FHFA indexes.Tis method allows economists to control orhome characteristics—the previous sale pricebeing considered an appropriate surrogate
or the hedonic in ormation. An obvious
limitation is the omission o sales o newhomes. Additionally, to maintain consis-tency, repeat sales indexes o en drop housesthat have undergone major improvements ordeterioration. Consequently, this method’scalculations require a large number o repeatsales, which can be problematic or nonmet-ropolitan areas and also during downturns.Finally, it has been shown that repeat saleswith larger time gaps in between transactions
have greater variance, leading some indexesto adjust their weight downward.wo median price indexes are notewor-
thy: the NAR index and the Census Bureaumedian index. Te ormer dates to 1968. Tedata come rom surveys o sales o existingsingle- amily homes rom NAR a liates.Te national median is calculated by value-weighting the median within each o the ourcensus regions in the country by the numbero single- amily homes in each region.
Te Census Bureau median index di ersrom the NAR index mainly in that the or-
mer covers new homes as opposed to exist-ing structures. Consequently, the Censusmedian index is typically higher than theNAR index (see Figure 1) since, historically,new homes have been higher-priced thanexisting homes. In terms o both indexes,prices have clearly allen since their peaksin 2006-07. However, the gap between thetwo has widened recently, largely due tothe steeper decline in the NAR index. Onepossible reason: Existing homes have seenan increase in oreclosures and short sales,placing downward pressure on the NARindex. Distressed sales are less o a concernin the market or new homes, and the Cen-sus median index has not allen as sharply.
Indexes o repeat sales are more com-monly cited than median indexes. TeFHFA index is published quarterly by theFHFA and goes back to 1975. Te FHFAalso publishes several other indexes, includ-ing regional, state, metropolitan, purchaseonly, average and median price indexes on amonthly and quarterly basis.
Standard & Poor’s publishes the Case-Shiller proprietary amily o indexes, whichincludes quarterly national, monthly 10- and 20-composite metropolitan area,
and individual metropolitan series.Te nal index is the monthly CoreLogicindex, a proprietary index published by CoreLogic and dating to 1976. Additionally,CoreLogic publishes a variety o indexesbased on property locations, price tiers, prop-erty types, loan types and distress levels.
Among the three major repeat salesindexes, the FHFA index is signi cantly di erent rom the other two. FHFA collectsdata rom con orming mortgages only (i.e.,
those securitized by Fannie Mae or Fred-die Mac).3 Te Case-Shiller and CoreLogicindexes, however, include all available arm’s-length transactions on single- amily homes,including sales nanced with noncon orm-ing mortgages—such as jumbo, Alt-A andsubprime. As a result, these indexes includesales o higher-priced homes (those nancedwith jumbo mortgages) and transactionswith more-volatile sales prices (those
nanced by Alt-A or subprime mortgages).Moreover, unlike the FHFA index, the Case-Shiller and CoreLogic indexes value-weighttransactions so that higher-valued homeshave greater e ect on the index. 4 A naldistinction is that the FHFA index includesre nances, whereas the Case-Shiller andCoreLogic indexes do not. 5
While the Case-Shiller and CoreLogicindexes are similar, they are di erent ontwo counts. In addition to value-weight-ing, the Case-Shiller series employs an
interval-weighting procedure that placesgreater weight on repeat sales with shorterintervals. Such a weighting scheme is notadopted by CoreLogic. Also, CoreLogic haslarger coverage because it includes mortgagedata in place o public records in states withnondisclosure laws. Tis helps it obtain abroader coverage by including some stateswith nondisclosure laws that are omitted inthe Case-Shiller index.
Figure 2 shows various repeat salesindexes.6 Notably, the FHFA index is at-ter than the other two indexes. First, theCoreLogic and Case-Shiller indexes placemore weight on higher-valued homes; so,i higher-priced homes have larger appre-ciations and, subsequently, larger depre-ciations, then these indexes will likely seelarger swings. Second, the FHFA index isless volatile because it does not include non-con orming loans. Combined, these actorscan help explain why the FHFA index is
atter than the other two series.Not surprisingly, the CoreLogic and
Case-Shiller indexes tend to move togetherbecause o their similar computation andincluded loan types. However, the Core-Logic index tends to be slightly higher thanthe Case-Shiller national index. Tis is pos-sibly due to the smaller weight on lengthierintervals between sales in the Case-Shillerindex. Stated di erently, the statistical pro-
cedure used in the Case-Shiller index likely mitigates the in uence o sales pairs withextreme price changes.
It is not always a act that home priceindexes move in tandem. It is not di cultto record instances where changes in homeprices di er in both direction and magni-tude. Tis is true, or example, o the FHFAand Case-Shiller indexes or the secondquarter o 2010. Te di erences in meth-odology and composition determine thebehavior o each index at di erent points in
time. Knowledge o individual index calcu-lation aids in understanding the observeddisparities among the indexes.
Rajdeep Sengupta is an economist and BryanNoeth is a research associate, both at theFederal Reserve Bank o St. Louis. See http:// research.stlouis ed.org/econ/sengupta/ or moreon Sengupta’s work.
E N D N O E S
1 Te FHFA house price index was ormerly titled the OFHEO index.
2 Tis methodology was developed by Bailey,Muth and Nourse and was later modi ed by Karl Case and Robert Shiller (1987, 1989).
3 See http://en.wikipedia.org/wiki/Non-con orming_mortgage
4 See Aubuchon and Wheelock.5 Te FHFA also publishes a purchase-only
index that excludes re nances.6 Note that the Case-Shiller i ndex is quarterly,
whereas the CoreLogic is monthly.
R E F E R E N C E S
Aubuchon, Craig P.; and Wheelock, David C.“How Much Have U.S. House Prices Fallen?”Federal Reserve Bank o St. LouisNational Economic rends, August 2008.
Bailey Martin J.; Muth, Richard F.; and Nourse,Hugh O. “A Regression Method or RealEstate Price Index Construction.” Journal o the American Statistical Association,December 1963, Vol. 58, No. 304, pp. 933-42.
Case, Karl E.; and Shiller, Robert J. “Pr ices o Single-Family Homes Since 1970: NewIndexes or Four Cities.” New England Economic Review, September/October 1987,pp. 45-56.
Case, Karl E.; and Shiller, Robert J. “TeE ciency o the Market or Single-Family Homes.” American Economic Review,March 1989, Vol. 79, No. 1, pp. 125 -37.
Lamppost banners around town bear the old-timeimage o a horse-drawn, two-wheeled cart with
a seated, mustached driver. Tus does Du Quoin,Ill., hitch its own wagon to its namesake event, theDu Quoin State Fair, long amed or harness racing.
Te air is the social, recreational and economicevent o the year, un olding over 10 days in late summer and peaking on Labor Day. It’s a something-
or-everyone a air, with auto and motorcycle races,horse shows, livestock and arm equipment exhibits,carnival rides, musical acts and ood. O course,there’s also harness racing, though the schedule wasreduced to three days rom ve last year, re ectingthe sport’s waning popularity, says the air’s man-
ager, John Rednour Jr.Attractions are spread out across the air’s 1,435
acres and its 77 buildings. Te centerpiece is a7,700-seat grandstand that looks out on a stageand the one-mile, lighted, circular racing track.An estimated 350 temporary jobs make the airbrie y the city’s largest employer. Upward o 300,000 people attend.
Local hotels are booked solid weeks in advance,with the over ow spilling out or miles around, saysStacy Hirsch, executive director o the Chamber o Commerce in Du Quoin (pronounced du-COIN ).
Te air boosts business or restaurants, gas sta-tions and stores, adds Judy Smid, president o theDu Quoin ourism Commission and proprietor o a downtown gi shop. Certi ed public accountantHarold Emling says the merchants tell him they getabout a month’s worth o their revenue rom the air
ypical o state airs, the Du Quoin eventcelebrates agriculture, historically a oundationo Southern Illinois’ economy, along with coal.Agriculture remains the Du Quoin area’s economicbase, says Je rey Ashauer, the city’s economic
By Susan C. Tomson
c o m m u n i t y p r o i L e
*
**
**
***
†
††
† °†
††
D Q oin/Perr Co nt , I . b bP pu t n f r C ty/C unty ......................6,109/22,350l b r F rce ....................................................Na/9,554Unemp yment R te .......................... Na/10.2 percent
Per C p t Pers n inc me.......................Na/$24,290 * U.S. Census Bure u, 2010 census
** BlS/HaVER, apr 2011, se s n y djusted*** BEa/HaVER, 2009
lARGEST EMPlOyERS
Gener C b e C rp. ............................................... 215He rt nd B k ng llC .............................................200M rsh Br wn ng H sp t ....................................190Du Qu n C mmun ty Un t Sch D str ct N . 300....187W -M rt ..................................................................110
† Se f-rep rted† † Reference USaGoV, inf gr up inc.
development consultant. Prospering romrecent high prices or their corn, wheatand soybeans, armers “come to town, buy new vehicles, put their money in the bank and go to Wal-Mart,” he says.
Coal, meanwhile, has taken its lumpssince the Clean Air Act dried up mar-kets or the state’s high-sul ur product.Most o the Du Quoin area’s coal mineswere shut down 20 or more years ago,and the resulting job losses account orPerry County’s chronically above-averageunemployment rate since, says DanielFulk, president o Du Quoin State Bank.
Te Great Recession? “Du Quoin hasbeen in a statistical recession or the last30 years,” he says. “We’re used to it.”Trough it all, the community has provedpersevering and resilient, qualities thatposition it to “survive very nicely in the
uture,” he adds.Rex Duncan, one o the our members
o the city’s elected governing commis-sion, also makes an upbeat case or thecity, based on its brick-and-mortar assets.Showing a visitor around, he points out
an Amtrak station with daily passengerservice to Chicago and New Orleans, an$18 million high school under construc-tion and a “basically new” hospital.
Marshall Browning Hospital dates to1922, but it has been almost completely rebuilt over the past decade at a cost o nearly $10 million. Emergency room,radiology department, pharmacy, labs,surgical suites, o ces, single-bed patientrooms—everything has been upgraded to
state-o -the-art. A physicians’ buildingand a 22-unit independent-living centerhave been added to the 19-acre campuson the edge o town.
Te hospital has an annual payroll o $6.5 million and, between its 25 bedsand extensive out-patient services, is ableto meet 75 percent o the community’shealth-care needs, says the chie executive,William Hu .
“We don’t have a lot o industry here,”observes Emling, while acknowledgingthat two o the city’s other leadingemployers happen to be manu acturers.
General Cable Corp. has been a xturein town under various names and ownerssince 1965, making insulated cable, espe-cially or electric utilities. Te plant hasthrived not only on good relations betweenmanagement and the eamsters-led labor
orce but also on the resulting exibil-ity to quickly change product lines, sayshuman resources manager Kathy Hanks.Although 20 production jobs were cutduring the recession, all have since beenadded back.
Heartland Baking, a commercial cookiemaker that started up in 2006 in a bakery shuttered by its previous owner, is oneo only two tenants in the city’s 90-acre,20-year-old industrial park. Te otheris MPP, an electroplating company thatmoved to the park rom Kansas City in 2000.
Besides state tax credits linked to thenumber o jobs created, Heartland andMPP each got a low-interest $400,000 loan
The Du Quoin State Fair c d w c . D
, c w d d d
c d c
As at any major county or state air, x ck
c c c .
The Reg n Ec n m st | www.st o is ed.org 19
photo by terry Jones
photo proviDeD by John Croessman,Du Quoin EvEning Call
unded by the ederal block grant program,with principal and interest repayable to thecity. Te city has used the income to makecivic improvements and other low-interestbusiness loans.
Ashauer says the park has proved a toughsell to out-o -town prospects or three rea-sons: U.S. manu acturing has been movingo shore, Illinois is perceived as business-un riendly and Du Quoin isn’t located onan interstate.
Tat “tough sell” has become a bit lesstough now that the city has decided to out tthe park with solar panels. Te installation,
nanced with $405,000 in ederal stimulusmoney plus $135,000 in city development
unds, is expected to be nished later thisyear. Te panels are expected to shaveelectric bills or park tenants by 10 percent.Given that promise, word o the park is nowspreading “like wild re,” Ashauer says.
As a urther inducement, the city in 2009classi ed the park and some nearby landas a tax-increment nancing ( IF) district,where any new or increased city real estatetaxes will be automatically reinvested.
Tis was the second o Du Quoin’s twoIF zones. Te original, set up two years
earlier, covers the dozen square blocks o the city’s Victorian-era downtown.
As the decades passed, Wal-Mart openedoutside o downtown and retailers gravi-
tated to malls. Te city’s core was showingsigns o wear and neglect. “Your downtownis like your home; i you don’t keep it up, it
alls apart,” says Mayor John Rednour Sr.,ather o the air manager.As one remedy, the city has set aside
$100,000 rom the downtown IF andmade it available in grants o up to $5,000to downtown owners or updating theirbuildings’ acades. Projects began this pastspring. In the interest o creating synergiesbetween downtown and the air, IF grants
are also available to downtown businessesthat make improvements designed to attractairgoers. One o these went to a restaurant
that added outdoor seating.Te air has been a source o pride, un
and dollars to its hometown since its birthas a private enterprise in 1923. It has goneon uninterrupted, even through its 1986takeover by the Illinois Department o Agri-culture, which also runs the larger state airevery August in Spring eld.
Independent audits, available rom thestate or the years through 2009, show theDu Quoin air losing money annually orthe previous decade on revenue averaginga little more than $1 million and expensesranging rom $1.5 million to more than$2 million.
But it’s un air to judge the air on thosenumbers alone, o cials say. For instance,while the audit shows a de cit o $863,288
or the 2000 air, the estivities spun o more than $8 million in economic bene tto Perry County, according to an analysisby the University o Illinois and the FederalReserve Bank o Chicago. A similar multi-plier e ect still applies, Rednour Jr. says.
Many air events are ree, and there is nogeneral admission charge, just a parking ee.While striving to keep the air “a ordable
or people to come with their kids and showthem a good time,” Rednour Jr. says he’salso nudging it toward break-even by cut-ting some expenses and raising some ees.
Separately, air managers several years agobegan a drive to lure paying events in theyear’s remaining 50½ weeks. With their 1,200electrically equipped campsites, the groundshave proved popular or RV rallies lasting aweek or more. Other non air money-makershave included bull riding, rodeos, monstertruck shows, charity events, weddings, picnics,
ea markets, demo derbies, horse shows, auto
races, motorcycle races and, o course, har-ness races. Annual revenue rom o -seasonbusiness has grown to between $650,000and $700,000, o cials say.
Tese extra events also boost the localeconomy, says Tomas Jennings, directoro the Illinois Department o Agriculture.As or the air proper, it’s “a good deal orthe community,” he says. “Te state sup-ports all o the communities in Illinois.Te air is our opportunity to supportSouthern Illinois.”
But or how long and how much? Untilthis year, the air’s uture was never indoubt, Emling says. With the state o Illi-nois acing a de cit o more than $9 billion
or scal 2012, all department budgets acecuts. And the air, like every other expense,will “have to work its way through the legis-lature,” Jennings says.
he macroeconomic environmentcontinues to improve, although the
pace o economic activity has been bumpy and somewhat lackluster. In particular, theunexpected slowing in real GDP growthduring the rst quarter (1.8 percent rom the
ourth quarter’s 3.1 percent) occurred againstthe backdrop o healthy increases in private-sector employment and a modest decline inthe unemployment rate.
As policymakers, businesses and con-sumers grapple with the lingering e ectso the nancial crisis and recession, someadditional risks have emerged. Chie amongthese are sharply higher energy prices andthe uncertainties stemming rom develop-ments in Europe, the Middle East and Japan.Still, most orecasters continue to believethat the economy will shake o the rst-quarter doldrums o unexpectedly highin ation and subpar output growth andwill soon transition to lower in ation rates
and a stronger pace o economic activity.1
He p Wanted
Among the most notable developmentso late has been the sharp rebound inmonthly private-sector payroll employ-ment. Although the pace o hiring slowedin May, private employment increased by 182,000 jobs per month over the rst vemonths o 2011. Average monthly gains intotal non arm payrolls were a bit smallerbecause state and local governments
reduced employment to help correct theirscal imbalances. However, the economy’sgrowth has not been brisk enough to bringabout dramatic reductions in the unemploy-ment rate, which remained quite elevated inMay (9.1 percent). Pro essional orecastersgenerally expect total non arm job gains toaverage about 190,000 per month throughthe rst hal o 2012, with the unemploy-ment rate slowly alling to about 8.25 per-cent by June 2012.
The Ret rn o Oiat $100 per Barre
Perhaps surprisingly, the rise in oil pricesand the resulting surge in average gasolineprices to near $4 per gallon nationally havenot yet derailed consumer spending orimpinged on planned capital expenditures by businesses. Te previous surge in oil prices,in 2007-2008, helped push the economy intorecession, but today’s dynamics are muchbetter: Equity prices are rising, real interestrates are lower, real household incomes arestrengthening, and housing construction andhousehold wealth are no longer plungingat a rapid rate. In addition, the rebound inglobal growth has bene ted many rms,especially manu acturers. Tis develop-ment, in conjunction with a weaker dollar,has kept U.S. exports expanding at a rapidclip. Relatively strong business expenditureson equipment and so ware are a key signalthat rms expect solid economic conditions
going orward.Te construction sector remains the y
in the ointment, as housing starts and newhome sales continue to linger near recordlows, and o ce and commercial constructionlanguishes. Moreover, house prices continueto dri lower because o the large number o unsold houses on the market and high ore-closure rates—although the latter have beentrending lower. Te growth o ederal gov-ernment outlays has also weakened becauseo the waning ederal stimulus program and
pressures to reduce the extraordinarily largebudget de cit.Te rise in oil prices and some o the
lingering uncertainties spawned by eventsoverseas have not shaken the con dence o
nancial markets either. Equity prices haverisen sharply since late August 2010, and theSt. Louis nancial stress index has returnedto its pre nancial-crisis levels. Improvingeconomic and nancial market conditionshave begun to increase the demand or bank
E N D N O E1 Re erences in this article to in ation are to “headline
in ation,” which actors in ood and energy prices.
loans by businesses, and consumer credithas started to rise modestly.
Infation Increases
Sharply higher energy prices, as well asrising ood prices, have pushed headlinein ation rates to levels last seen during the2007-08 oil price shock. Over the past year,the CPI rose by 3.4 percent. A key worry associated with an oil shock (or higher oodprices) is the impact that “pass-through”e ects may have on prices o non ood andnonenergy goods and services. I long-termin ation expectations are viewed as low andstable and i monetary policy is viewed ascredibly committed to long-term price sta-bility, then these pass-through e ects tendto be modest and temporary.
Accordingly, most economists and Federal
Reserve policymakers view the sharp rise inin ation as a temporary deviation rom a lowand stable in ation environment. As long asthis expectation persists, the unemploymentrate remains high and the pace o growthuneven, most orecasters and nancial mar-ket participants believe that the Federal OpenMarket Committee will maintain its existing
ederal unds rate target o 0 to 0.25 percent-age points or the remainder o 2011—andmaybe into the rst hal o 2012.
Kevin L. Kliesen is an economist at the FederalReserve Bank o St. Louis. See http://researchstlouis ed.org/econ/kliesen/ or more on his wo
The Eighth Federal Reserve District c d c w c c d d
l r ck, l , m d s . l
Hispanics P a Di erent Ro e
in District’s Growth than in Nation’sBy Rubén Hernández-Murillo and Christopher J. Martinek
he U.S. Census Bureau recently releasedthe 2010 redistricting data or the nation.
Tese data are the rst to provide local-levelin ormation on population, race/ethnicity,age and housing unit counts rom the2010 census. Aside rom helping de ne
congressional district boundaries, the datareveal interesting trends over the pastdecade across various demographic groups.One trend that has received a lot o atten-tion is the dramatic growth o the Hispanicpopulation, which in 2010 represented 16.3
percent o the nation’s population. 1 Tedemographic trends in the Eighth FederalReserve District in terms o populationgrowth by racial and ethnic categories werequite di erent rom the national trends. 2
Te table provides a snapshot o population
United States and EighthDistrict Comparison
2000Population
2010Population
Changesince 2000
PercentageChange
HispanicContribution
to Growth
Non-HispanicWhite AloneContribution
to Growth
Non-HispanicBlack AloneContribution
to Growth
Non-HispanicAsian AloneContribution
to Growth
Non-HispanicOther Single Race
Contributionto Growth
Non-HispanicMultiple RaceContribution
to Growth
United States 281,421,906 308,745,538 27,323,632 9.7% 5.4% 0.8% 1.3% 1.5% 0.2% 0.5%
growth by race and Hispanic origin in theU.S. and the Eighth District. Te top panelsummarizes di erences in rural and urbanareas, while the bottom panel illustratespopulation trends across metropolitan areasin the Eighth District.
Overa Pop ation Growth
Between 2000 and 2010, the nation’s popu-lation grew by 9.7 percent to 308,745,538.About 56 percent o the growth in U.S. totalpopulation was accounted or by individualswho identi ed themselves as Hispanic orLatino (5.4 out o 9.7 percent). In the EighthDistrict, total population between 2000 and2010 increased by 6.2 percent to 14,569,665.Hispanics represented 3.6 percent o the Dis-trict’s total population. Although the contri-bution to growth o the Hispanic popula-tion was the largest among all groups, itaccounted or only about a third o totalpopulation growth (2.0 out o 6.2 percent).Almost 50 percent o the total growth inthe Eighth District was accounted or by thecombined growth o non-Hispanic indi- viduals who identi ed themselves as non-Hispanic white alone or non-Hispanic black alone (1.7 and 1.3, respectively, out o 6.2percent). Growth in the non-Hispanic Asianpopulation was the second largest contribu-tor to national population growth, represent-ing about 15 percent o overall growth (1.5
out o 9.7 percent), but in the Eighth District,the population growth o non-HispanicAsians accounted or only about 8 percent o overall growth (0.5 out o 6.2 percent).
R ra and urban Growth
Although Hispanics’ contribution to over-all growth was less dramatic in the EighthDistrict than in the nation as a whole, break-ing up total population across urban andrural counties reveals that Hispanic popula-tion growth was a more important contribu-
tor to rural population growth in the EighthDistrict than in the nation. Tis distinctionis important because the Eighth District ismore rural than the nation as a whole.
Te 2010 census indicates that 39.1 percento the District’s population lives in ruralcounties, while only about 17 percent o thenation’s population lives in rural counties. 3 Te growth in rural population o the nationwas 4.4 percent, while the growth in urbanpopulation was 10.8 percent. Te population
in rural counties o the Eighth District grewby 1.6 percent, while population in urbancounties grew by 9.4 percent. 4
In terms o contributions to growth,Hispanic population growth accounted orabout 55 percent o the nation’s populationgrowth or both rural and urban counties(2.4 o 4.4 percent in rural counties and 6 o 10.8 percent in urban counties). In contrast,Hispanic population growth accounted or75 percent o relatively modest rural popula-tion growth in the Eighth District (1.2 o 1.6percent) and slightly more than 25 percent o urban population growth (2.5 o 9.4 percent).
MSA Pop ation Growth
Across the Eighth District’s metropolitanstatistical areas (MSAs), with the exceptiono Pine Blu , Ark., population increased inevery metropolitan area rom 2000 to 2010.Fayetteville-Springdale-Rogers, Ark.-Mo.,led the District MSAs with a 33.5 percentpopulation growth. Te largest contribu-tions to growth in this location came romthe Hispanic population, with about 34 per-cent o overall growth (11.6 o 33.5 percent)and rom non-Hispanic white individuals,with about 47 percent o overall growth(15.9 o 33.5 percent).
Population growth in most o the DistrictMSAs was driven predominantly by growthin the non-Hispanic white population. Te
exceptions were Memphis, enn.-Miss.-Ark.;exarkana, exas-Ark.; Jackson, enn.;
and most notably, Pine Blu , Ark., wheredecreases in the non-Hispanic white popula-tion subtracted rom overall growth. Incontrast, growth in the St. Louis, Mo.-Ill.,and Jonesboro, Ark., areas can be predomi-nantly attributed to growth in the non-Hispanic black population. Growth in thenon-Hispanic Asian population also made upa signi cant proportion o total populationgrowth in the St. Louis MSA. Fort Smith,
Ark.-Okla., and Owensboro, Ky., moreclosely resembled the national trend o His-panic population growth accounting or thelargest share o total population growth.
Rubén Hernández-Murillo is an economist and Christopher J. Martinek is a research associate,both at the Federal Reserve Bank o St. Louis. Seehttp://research.stlouis ed.org/econ/hernandez/ or more on Hernández-Murillo’s work.
E N D N O E S
1 Te census collects race and Hispanic originin ormation in accordance with the U.S. O -
ce o Management and Budget’s (OMB) 1997Revisions to the Standards or the Classi ca-tion o Federal Data on Race and Ethnicity,which prescribe that race and Hispanic originbe considered distinct concepts necessitatingthe separate questions.
2 For the purposes o this ar ticle, we compareHispanics with individuals who reportednon-Hispanic origin and only one race (white,black or Asian) to orm mutually exclusivecategories.
3 Urban counties, here, are de ned as thosemaking up part o a census-designatedmetropolitan statistical area.
4 Some counties o MSAs listed in the lowerportion o the table are located outside o theDistrict and are not included in the gurespresented in the upper portion. For example,in the Fort Smith, Ark.-Okla., MSA, SequoyahCounty, Okla., is located outside o theDistrict. Similarly, some counties located inMSAs considered outside the District and notincluded in the lower portion o the table areincluded in the tabulation or the upper por-tion o the table, or example, Greene County,Ind., in the Bloomington, Ind., MSA.
C E N S U S C H A N G E S
Unlike previous censuses, the 2010 censusdid not include a “long orm” questionnaire.Previously, the long orm was given to roughly one in six households to gather in ormation onsuch things as educational attainment, income,housing costs and other socio-economic char-acteristics o the population. (Te long ormcontinues to be administered every year as parto the American Community Sur vey.)
One o the reasons or eliminating the longorm was to improve return rates. Te mail
participation rate or the 2010 census was 74 per-cent o occupied households, the same rate thatwas achieved or the 2000 census short orm.However, when the elimination o the long ormis actored in, a larger portion o questionnaireswas returned in 2010.
Te Census Bureau makes an attempt to ol-low up with households that do not respond by mail; the bureau will ca ll, visit the household orcontact neighbors and building managers. As alast resort, the bureau will impute counts usingstatistical models that re ect the characteristicso the neighborhood. By the time all the meth-ods o lling in missing orms are exhausted,the bureau determines the proportion o recordsthat provide usable in ormation. Last year, thisproportion was 99.62 percent, slightly higherthan the 2000 proportion o 99.43 percent.
In addition to the response rates, the bureauconsiders several other measures o accuracy o the data-collection process. One o the mostimportant post-census process indicators isthe Census Coverage Measurement survey, aquality-check survey o 300,000 households.Results rom this survey wi ll be matched tocensus responses to estimate overcounts andundercounts by geography, ethnicity, race, gen-der and age. Te bureau will publish the resultsnext year but will not revise ex isting populationcount estimates.
Based on a popular index, racial segrega-tion decreased in the Eighth District’s
our major metropolitan areas between 1970and 2000. Tis decline was not particularto the Eighth District; or example, a similardecline occurred in Chicago.
o help explain what happened, we cre-ated a simple way to decompose the declinein the index; by doing so, we ound thatthe decline can be explained by opposing
orces that are the same in all metro areas.Te orce that lowered the index o segrega-tion was an increase in racial integration inhistorically highly black and highly whitecommunities. Te orces that partly o setthis decrease were the suburbanization o thewhite population into new, highly white com-munities and, to a lesser extent, the increased
segregation in communities that experienced“tipping” rom highly white in 1970 to highly black in 2000.
The Basics o O r St d
Racial segregation exists in a city to theextent that people o di erent races donot share the same areas. 1 Di erent typeso areas can be analyzed, such as blocks,neighborhoods or counties. For this article,we documented the extent and evolutiono black/white segregation across census
tracts o the Eighth District between 1970and 2000.2 Although 1970 is a good startingpoint (since it was the rst decennial censusyear a er the Civil Rights Act o 1964), we
ocused on the 1970-2000 period mainly because there exist adequate data or it.
Te data we used come rom the Neigh-borhood Change Database (NCDB). 3 Tisdataset is built by trans orming the origi-nal Census Bureau data in such a way thattract borders do not vary between 1970 and
2000.4 Using it, we could observe segrega-tion changes within xed plots o land. (Data
rom the 2010 census are not yet availablein the NCDB ormat.) We used the Indexo Dissimilarity (IOD), a popular measureo segregation among sociologists andeconomists, because it has a straight orwardinterpretation.
The Index o Dissimi arit
Te IOD varies rom zero to 100 percent.An IOD o 90 implies that at least 90 percento one o the two groups (in this case, eitherblack or white) would need to move to adi erent neighborhood to make all neighbor-hoods end up with the same racial mix.
Consider a dessert party in which twobuckets o vanilla ice cream and one bucket
o chocolate ice cream are to be served.o serve all guests with the same vanilla-
chocolate mix, each guest would need tobe served two scoops o vanilla with eachscoop o chocolate. I each bucket contains100 scoops, all one ends up doing is serving1 percent o the total amount o vanilla icecream together with each 1 percent o thetotal chocolate ice cream. Te IOD captureshow ar the party is rom the homogeneousdistribution by comparing the percentageso the total chocolate and vanilla ice cream
served onto each plate. For example, a platethat contains 5.7 percent o the chocolateice cream and 1.3 percent o the vanilla icecream contributes (5.7% – 1.3%) to the IOD(i.e., 4.4 percentage points). Adding up thecontributions rom all plates with excesschocolate gives the total index. (Te calcula-tion is identical i we consider plates withexcess vanilla instead.) When the percent-ages are equal on all plates, the index is zero.When no plate contains both avors, the
index is 100 percent— ull segregation.For a concrete example, consider St. Louis
in 1970. In that year, the population o St. Louis was 2,071,043. O those, 375,090persons were black and 1,688,491 were white.St. Louis as a whole was 18.2 percent black.
Te le panel o the diagram summarizessegregation in St. Louis by joining all tractsthat were more than 18.2 percent black intowhat we call the “highly black” (HB) area andby joining all tracts that were less than 18.2percent black into what we call the “highly white” (HW) area. Te diagram shows that94.2 percent o the black persons in St. Louislived in HB tracts while only 10.6 percento the white persons lived in those tracts.(Recall that these two percentages would haveneeded to be equal or the neighborhood to
have been exactly 18.2 percent black.) Onehypothetical way or the HB area to become
ully integrated would be to reduce its per-centage o blacks in that area to 10.6 percent,which would be equal to the percentage o whites in that area. o achieve this reduction,the equivalent o 83.6 percent o all black people in St. Louis would have needed to moveout o the HB area. I this amount o black people would have moved into the HW area,the percentage o all black persons living inthe HW area would have risen rom 5.8 per-
cent in the diagram to 89.4 percent—exactly equaling the percentage o all white personsliving in the HW area. Tere ore, this move-ment would have su ced to achieve per ectintegration in HW and also in HB areas.
In summary, 83.6 percent o all black persons in St. Louis would have needed tochange neighborhoods in 1970 in order tomake all areas ully integrated. Tis percent-age was the IOD or St. Louis in 1970. Tisexercise could be repeated with the white
24 The Reg n Ec n m st | J 2011
Black/White Segregationin the Eighth District:A Look at the Dynamics
population moving out o HW areas, and theresulting IOD would be unchanged.
IOD’s Change over Time
In 1970, the IOD in District metropoli-tan statistical areas (MSAs) was very high,ranging rom 73.3 percent in Little Rock to83.6 percent in St. Louis, while it was slightly above 90 percent in Chicago. Te IOD ell orall MSAs in our table between 1970 and 2000.Te largest declines happened in Louisville(20 percentage points) and Little Rock (15percentage points), while Chicago, St. Louisand Memphis observed milder declines(approximately 12 percentage points).
o get some notion as to why the IOD ellin all o our MSAs, consider the right panel o
the diagram. Te diagram shows how citieschange between two points in time—say 1970and 2000. In 1970, the city is representedby solid lines, and area types 1, 2 and 3 areHB, while 1*, 2* and 3* are HW, just like inthe le panel o the diagram. In 2000, thecity is represented by dotted lines. Each arearepresents neighborhoods that experienceddi erent kinds o changes between 1970 and2000. We can name each kind o changeusing popular terminology:
White Resegregation: racts that stay HW,represented by area 1.
Black Resegregation: racts that stay HB,represented by area 1*.
ipping Black to White: racts that switched
note: W c s . l “i s 1970” , c d msa. W d cd c “C w 1970 d 2000” . s c c ’ d d .
The Reg n Ec n m st | www.st o is ed.org 25
Diagram o a Segregated Cit and Its Change over Time
Highly Black Tracts (St. Louis)t p : 531,772r c m x: 66.4% ck
p c b ck p : 94.2%p c W p : 10.6%
Highly White Tracts (St. Louis)t p : 1,531,809
r c m x: 1.4% ckp c b ck p : 5.8%p c W p : 89.4%
i n i t i a l s i t u a t i o n i n 1 9 7 0 C h a n g e b e t W e e n 1 9 7 0 a n D 2 0 0 0
1*4*2* 3*
Little Rock Louisville Memphis St. Louis Chicago
Index o Dissimilarity 1970 73.33 81.42 82.31 83.58 90.17
Index o Dissimilarity 2000 58.29 60.77 70.23 71.95 77.74
Change 1970 to 2000 –15.05 –20.66 –12.07 –11.64 –12.43
notes: e c d c d . n d c . n c dd x c c d .
Decomposition: Contribution to Change by Each Type o Tract (Percentage Points)
Black Resegregation (1) –15.0 –14.3 –22.4 –15.1 –18.2
Tipping B to W (2) 0.0 –0.1 0.1 0.0 0.0
Black Depopulation (3) –2.6 0.0 –0.3 –0.1 –0.1
Black Suburbanization (4) 0.0 0.1 0.7 0.1 0.3
White Resegregation (1*) –12.7 –12.0 –6.3 –4.3 –5.7
Tipping W to B (2*) 2.8 0.2 0.6 1.1 5.3
White Depopulation (3*) –0.9 –1.1 0.0 –0.2 0.0
White Suburbanization (4*) 13.3 6.5 15.4 6.8 6.0
TOTAL –15.05 –20.66 –12.07 –11.64 –12.43
Index o Dissimi arit in 1970 and 2000, Eighth District and Chicago (Percent)
4
32
1
E N D N O E S
1 Te U.S. pattern o racial residential segrega-tion has been studied by economists sincethe mid-20th century, ollowing the seminalworks o Gunnar Myrdal and, later, TomasSchelli ng. Sociologis ts have also made impor-tant contributions to the measurement andtheory o racial segregation. For an overviewo segregation measurement, see ww w.census.gov/hhes/www/housing/resseg/app_b.html
2 Census tracts are small units o land delineatedby the Census Bureau. Tese units subdivide acounty and usually contain between 2,500 and8,000 people.
3 ract level data come rom the NeighborhoodChange Database (NCDB) by Geolytics Inc.Te database contains tract-level populationcounts rom the 1970, 1980, 1990 and 2000U.S. decennial censuses.
4 Te Census Bureau rede nes tract boundariesor each decennial census.
5 In this population count, we only considerblack and white population. We also considertracts with population density o ewer than100 people per square kilometer as empty andnormalize their population to zero.
6 Note that an empty tract contains zero percento each o the populations, so that it contri-butes 0 percent to the Index o Dissimilarity.Te change in segregation in these areas isthe new level o segregation (zero) minus theold level.
R E F E R E N C E S
Schelli ng, Tomas C. “Dynamic Models o Seg-regation.” Journal o Mathematical Sociology,May 1971, Vol. 1, No. 2, pp. 143-86.
Myrdal, Gunnar. “An American Dilemma: theNegro Problem and Modern Democracy.”New York: Harper & Brothers, 1944.
Massey, Douglas; and Denton, Nancy. “TeDimensions o Racial Residential Segregation,”Social Forces, December 1988, Vol. 67, No. 2,pp. 281-315.
u . s . a g r i C u lt u r a l t r a D e Fa r m i n g C a s h r e C e i p t s
06 07 08 09 10 11
80
60
40
20
0
NOTE: Data are aggregated over the past 12 months.
Exports
Imports
AprilTrade Balance
B I L L I O N S
O F D O L L A R S
06 07 1108 09 10
190
170
150
130
110
90
NOTE: Data are aggregated over the past 12 months.
FebruaryCrops Livestock
B I L L I O N S
O F D O L L A R S
C i v i l i a n u n e m p l o y m e n t r a t e i n t e r e s t r a t e s
06 07 08 09 10 11
11
10
9
8
7
6
5
4
P E R C E N T
May
06 07 08 09 10 11
6
5
4
3
2
1
0
10-Year Treasury
Fed Funds Target
May1-Year Treasury
P E R C E N T
NOTE: On Dec. 16, 2008, the FOMC set a target range forthe federal funds rate of 0 to 0.25 percent. The observationsplotted since then are the midpoint of the range (0.125 percent).
i n F l a t i o n - i n D e X e D t r e a s u r y y i e l D s p r e a D s r at es on Fe De ra l F un Ds Fu tu res on se le Ct eD Dat es
3.02.5
2.01.51.00.50.0
–0.5–1.0–1.5–2.0–2.5
NOTE: Weekly data.
5-Year
10-Year
20-Year
P E R C E N T
June 10, 2011
07 08 09 10 11 June 11 July 11 Aug. 11 Sept. 11 Oct. 11 Nov. 11
0.22
0.19
0.16
0.13
0.10
CONTRACT MONTHS
P E R C E N T
1/26/11
3/15/11 6/16/11
4/27/11
r e a l g D p g r o W t h C o n s u m e r p r i C e i n D e X
06 07 08 09 10 11
8
6
4
2
0
–2
–4
–6
–8
NOTE: Each bar is a one-quarter growth rate (annualized);the red line is the 10-year growth rate.
P E R C E N T
Q1
06 07 08 09 10 11
6
3
0
–3 P E R C E N T C H A N G E F R O M
A Y E A R
E A R L I E R
May
CPI–All Items
All Items Less Food and Energy
e c o n o m y a t a g L a n c e
rom HB to HW, represented by area 2.ipping White to Black: racts that switched
rom HW to HB, represented by area 2 *.Depopulation: racts that became vacant,
represented by areas 3 and 3 *.Suburbanization: racts that were empty
in 1970 but became populated by 2000, repre-sented by areas 4 and 4*.
For any city, each area described by theright panel o the diagram contributes tothe change in the IOD over time. Tiscontribution depends on the size o the areaand on the change in segregation withinthe area. Tere ore, we can decomposetime changes o the IOD by calculatingthe portion that accrues to each area. Tetable presents this decomposition, and wedescribe its contents below.
Both White Resegregation and Black Resegregation had large negative e ects onthe IOD. Tis means that although many tracts stayed HB or HW between 1970 and2000, these types o tracts became more mixed.
ipping White to Black appreciably helpedto increase the IOD in Chicago (5.3 percent-age points) and Little Rock (2.8 percentagepoints). Tis implies that the tipping tractsbecame at least as segregated a er becomingHB as they were when HW. ipping Black toWhite did not have a large e ect on the index
in any MSA.Depopulation o HB tracts reduced theIOD in Little Rock by 2.6 percentage points,while the e ect in other MSAs was below onepercentage point. Tis means that the tractsthat were HB in 1970 and were empty or very sparsely populated by 2000 were highly seg-regated in 1970. 6 In contrast, Depopulationo HW tracts did not appreciably change theIOD. Suburbanization into new HB tractsdid not impact the index appreciably, exceptin Memphis, where it increased the index by 0.7 percentage points. In contrast, Suburban-ization into HW tracts had a large positivee ect on the index in all MSAs, with the larg-est e ects in Little Rock and Memphis.
Alejandro Badel is an economist and Christo- pher J. Martinek is a research associate, bothat the Federal Reserve Bank o St. Louis. Seehttp://research.stlouis ed.org/econ/badel/ or more on Badel’s work.