Why Are We In Are Cession

Why are we in a recession?

The Financial Crisis is the Symptom, not the Disease!

Ravi Jagannathan

Kellogg School of Management Northwestern University Based on: p and Ernst Schaumburg g Joint work with Mudit Kapoor ©2009 Ravi Jagannathan

The Great Recession  US Corporate Equities  1999: $19.4 $19 4 Trillion (2.1 (2 1 GDP)  2008: $15.2 Trillion (1.1 GDP)

 US HH Net Worth  1999: $42.1 Trillion (4.4 GDP)  2008: $51.7 Trillion (3.6 GDP)

 Peak Unemployment  1999: 4.4% 4 4%  2008: 7.2%  June 2009: 9.5%

Folk Wisdom • What caused the Global Recession? – A. Financial Crisis

• What caused the Financial Crisis? – B. Easy Credit & Lax Regulation

• What caused Easy Credit & Lax Regulation? – C. Savings Glut, Too much money chasing too few opportunities

• What caused the Savings Glut? – D. D Too much saving in Asia Asia, and too little in the US

• Why Too Much Saving in Asia? – E. Asians like to save

• So what can we do to get out of the recession? – F. Asians should save less; Americans should save more

• Is that all that is needed – will it work?

A Deeper Driving Force? • What is wrong g with this logic? g • It is misleading to think that the causality is from – Too much savings in Asia, E → D → C → B → Financial Crisis, A → Global Recession

• All these phenomena are closely interlinked and causality flows both ways • There are deeper underlying driving forces – Changes in Geo-Political-Organization of Countries – Major Technological Innovations

Global Perspective: Major Shocks to World Economy • Geo-Political-Organization g – The opening of China – The opening of India

• Technological g Innovations – Communication & Transportation in the 90’s

Impact on Developed World’s Economy • Chinese factories can compete directly with US factories • Workers from India can directly compete with workers in the US • Workers in Developing World can participate in the Developed World’s labor market without moving • Huge increase in the Worlds’ labor supply in a very short time period

Impact on Developed World’s Economy … • I am going to argue that the inability of existing Financial and Legal Institutions to cope with this change is the reason for the recession – Large increase in employment in Developing Economies based on exports •

•

Inability to absorb savings through increased domestic investment and consumption  Inadequate national financial markets Currency controls motivated by national objectives

– Inability of US economy to adjust to perverse incentives caused by huge money inflows •

Institutional incentives, checks and balances in place turned out to be inadequate • Set the stage for the recession • The financial crisis was the first symptom

Rest of the Talk • • • • • • • •

Global trade and the savings glut Where did the savings go? US Household Behavior Savings glut or glut in labor supply or both? Financial engineering and the housing bubble What makes a housing g bubble different The way forward Final thoughts

Global Trade in Goods and Capital Flows

Global Trade and Savings Glut

Global Trade – Benefits • Increased living standards in the developing world • Cheaper goods and services in the developed world – Low inflation in developed world throughout 1982 – 2007

• 1982 98 – 2008: 008 Emerging e g g eco economies o es in Asia sag growing o ga at +7% pa

Global Trade and Savings Glut

Global Trade – Savings Glut • 1997 Asian financial crisis • BRIC + NIAC (+ Middle East Oil exporters) – Obsessed with building dollar reserves (through exports) • Dollar being the reserve currency

– Huge Current Account Surplus (Savings)

Global Trade and Savings Glut

Global Trade – Savings Glut … • Combined Current Account Balance of BRIC NIAC ME ($ billion): BRIC+NIAC+ME billi ) 1996: 2000: 2004: 2007:

+$ 4 + $ 149 + $ 318 + $ 798

• US Current Account Balance 1996 2000: 2004: 2007:

- $ 125 - $ 417 - $ 640 - $ 739

• Huge US Current Account Deficit – US $ should have depreciated p • say, relative to Yuan Global Trade and Savings Glut

Yuan per US$ did not Rise!

Global Trade and Savings Glut

Capital Flows • Huge Capital Flows into the US during the 21st century – To balance the huge current account deficit – Especially from China

Global Trade and Savings Glut

Where did the money go? • China emerged as a major creditor of the US • Chinese holdings of US assets • Initial Flow (1994 holdings) – $72 billion in Treasuries – $20 billion in Agencies – Lowering L i iinterest t t rates t on treasuries t i (in (i partt due d tto declining d li i deficits)

• Subsequent Flow (2007 holdings) – $466 billion in Treasuries – $376 billion in Agencies and some in corporate debt and equities – (Treasuries + Agencies): 0.25×GDP_China Where did the money go?

Change in China Holdings of US Assets *

* Corporate = non-agency non-government debt

Subsequent Flow into US … • To accommodate the money inflow investment banks set up their own pools of private label non-conforming mortgages providing investors the desired higher yields with (hidden, ill understood) risk • US Policy to promote increased home ownership played l d a major j role l g market • That funneled the moneyy flow into US housing – Home mortgages (indebtedness) went up as Current Account Deficit became large

Where did the money go?

Current Account Deficit & Household Debt/HH • Current Account Balance and US Household Debt

Where did the money go?

US Households • Why did US households consume more leading to the huge increase in Current Account Deficit? • Why Wh did the they increase their borro borrowing ing b by so m much? ch? • I am g going g to argue g that US households behaved rationally!!

US Household

Household Consumption & Wages • Private Consumption to GDP Ratio – 2000: 68.7% – 2007: 70.3%

• Wages, salaries, benefits, social security, and proprietors income to GDP Ratio – 2000: 66 66.3% 3% – 2007: 64.2%

• Why did consumption increase even though Wages remained flat? – Households felt wealthier! US Household

Stock and Housing Markets • Stock Market, S&P500 Index, remained flat – 2000 Q1: 1,499 – 2007 Q1: 1,421

• Housing prices increased by 86% • S&P/Case – Shiller Home Price Index – 2000 Q1: 100.77 100 77 – 2007 Q1: 186.07

US Household

Stocks: S&P 500 Index Values

US Household

Housing: S&P/Case – Shiller Index Q1 to Q1 200.00 180.00 160.00 140.00 120.00 100.00 80 00 80.00 60.00 40.00 20.00 0.00 1999

2000

2001

2002

2003

2004

US Household

2005

2006

2007

2008

Home Values/Equity per Household • Household Real Estate Value (per household) – – – – –

1980: $36,437 2000: $108,633 2007: $172,197 1980 to 2000: +$72,916 2000 to 2007: +$63,558

• Household Home Equity (per household) – – – – –

1980: $24,967 2000: $62,590 2007: $81,315 1980 to 2000: +$37,623 (52% of Household Real Estate Increase) 2000 tto 2007 2007: +$18,725 $18 725 (29% off H Household h ld Real R lE Estate t t IIncrease))

• So, they took out some of that wealth out to consume, increasing indebtedness (leverage). US Household

Mortgage/Home Value (Leverage)

US Household

After 2000: Sharp Rise in Household Debt

Financial Engineering

A Model of Household Consumption Choice

• Based on Rubinstein (1974) & Campbell and Cocco (Journal of Monetary Economics, 2007) Y Year t year change to h i h in household h ld consumption ti = 683 + 0.71×(change in wages and salaries + proprietors’ income) + 0.084×(change in household home equity) - 0.000×($ change in other household assets) (almost no effect) + error US households consume 71% of their salaries & in addition consume 8.4% of any increase in real estate wealth.

US Household

Predicted vs Actual Change in Household Consumption

×

US Household

Household Consumption: Recap

• US households consumed substantially more than their income during this 21st century • Th They did so b because th they ffelt lt wealthier lthi as h home prices i went up • That helped them borrow using home mortgage loans • To T follow: f ll ‒ Why did home prices go up? ‒ What is the relation to the current account deficit? US Household

CA Deficit and Capital Flows

• A $100 increase in GDP per US household • $22 spent on imports • Underlying forces that trigger the $100 increase in GDP, also increase exports by $14 • $8 current account deficit • $8 flows back into the US as capital account flows (for convenience, say from China) US Household

CA Deficit – Capital Flows – Home Prices • In early years (late nineties and early this century,) most of the $8 invested in US Treasuries • Made Treasury yields drop and less attractive • Made Mortgage related debts more attractive • Subsequent capital account flows went to mortgage backed debt • Resulting in lower mortgage interest rates, rates and easier availability of loans • Led to increased demand and higher housing prices • Feeds back through wealth effect leading to increased consumption, imports, and CA deficits US Household

Home Prices – CA Deficit – Feedback Effect

US Household

Testing for Feedback Effect • When CA Deficit Increases • Pool of outstanding home mortgages should increase (due to reduction in mortgage rates) $ Change in pool of home mortgages = +281 + 1.1 1 1 (CA Deficit) D fi it) - 0.26×(Change in level of US Treasury Debt) + error

US Household

Testing for Feedback Effect … • Home values increase when ‒ Pool of outstanding g home mortgages g g increase;; ‒ Mortgage interest rate decreases ‒ CA Deficit increases •

Other channels, viz. ease with which mortgage loans can be taken

$ Change in Household Real Estate Value = -1,063 - 181×(Mortgage interest rate) + 1.21×(Change in size of outstanding mortgage pool) + 2.91×(CA Deficit) + error

US Household

Back to Labor Supply Shock • What does labor supply shock have to do with all this?

Labor Supply Shock

Back to Labor Supply Shock

• Most of the post WWII miracle growth countries attained that growth through exports • Japan, Taiwan, S. Korea, Hong Kong, Singapore • More recently, China’s, C and to a lesser extent, India • However, population of China (and India) is huge • China: 1 1,300 300 million, million India: 1 1,200 200 million • Japan + Taiwan + S. Korea + Hong Kong + Singapore: 212 million

Labor Supply Shock

Back to Labor Supply Shock …

• China’s China s growth during the last two decades has been most impressive, and export driven (Offshoring) • China’s share of World GDP • • •

1980: <1% 1980 1% 2007: ~6% On track to overtake Japan by 2011 as the second largest economy

• China’s Exports/GDP • 1980: <6% • 2007: 2007 ~29% 29%

Labor Supply Shock

A Stylized Model of Offshoring • • • •

Two countries,, Rich (R) ( ) and Poor (P). ( ) Capital available (K): K = 1 in R and 0 in P Labor available (L): L = 1 in both R and P T (Leontief) Two (L ti f) technologies: t h l i ‒ ‒ ‒ ‒

Technology in R needs 1 of K and 1 of L and produces 2 units. Technology in P needs 0.1 of K and 0.8 of L and produces 0.51 unit. R’ ttechnology R’s h l can produce d either ith consumption ti or capital it l good d P’s technology can only produce consumption good

• Suppose the two countries cannot trade ‒ R employs K = 1, L = 1, produces 2, pays 1 to L (wage = 1), returns 1 to K (rate of return on capital = 0%); net output = 1.0 ‒ P does not have any capital, and so output = 0 ‒ Total net output of R and P = 1.0 10 Labor Supply Shock

A Stylized Model of Offshoring .. • Suppose pp some capital p can move between R and P •

R lends 0.1 unit of output (capital) to P and asks for 0.11 units to be returned (10% rate of return on capital)

• P uses K = 0.1; L = 0.8; and produces 0.51 unit of output; and pays 0.11 units to R and 0.40 units to L. ‒ Wages in P = 0.40/0.80 = 0.5; 20% unemployment

• R’s Return on capital rented out to P is 10% • Both countries are strictly better off by trading in capital • Total net output of R and P = 1 + 0.4 = 1.41

Labor Supply Shock

A Stylized Model of Offshoring … • Now suppose pp in addition, country y R can set up pp plants in P using its technology and employ P’s labor • Suppose R can train and employ the unemployed 0.2 (20%) of country P P’s s labor at lower wages compared to wages in R • No labor mobility across countries

Labor Supply Shock

A Stylized Model of Offshoring …. • R will move 0.2 of its capital to P, and keep 0.8 of its capital in R • •

Employ 0.2 L of labor in P, and 0.8L of labor in R Output of 1.6 in R: • return 0.8 to K in R (0% rate of return on capital) • pay 0.8 to L in R (wage remains at 0.8/0.8 = 1 unit)

•

Output 0.4 in P: • return 0.22 to K in R (10% rate of return on capital) • pay 0.18 to labor in P (wage = 0.18/0.2 = 0.9 > 0.5 for labor employed using P’s technology by local firms)

• Total output in R and P remains same at 1.41 • 20% unemployment in R and full employment in P • Those in P working for R strictly better off • •

The 0.2 unemployed p y in P were employed p y and received 0.18;; those with capital in R received 0.02 more. Those unemployed (0.2 or 20%) lost 0.2

A Stylized Model of Offshoring ….. • This model does not seem to capture what happened to the developed p countries of the West when Japan, p , Taiwan, Korea … developed through exports • Western economies also gained • What is wrong with the model? ‒ We assumed that the 0.2 of labor in R replaced by 0.2 of labor in P will remain idle. But they can be redeployed in other productive activities •

That will increase output in R. A Win-Win situation.

‒ If consumption goods become cheaper the labor in R may still be better off in real terms

Labor Supply Shock

A Stylized Model of Offshoring …… • Redeployment of labor in R takes time • Can become an issue if the magnitude of labor redeployment involved are large within a very short time ‒ I think that this is the case now with China’s China s and India India’s s export (offshoring) driven growth drive

‒ To summarize, summarize 1) big increase in labor supply in the offshoring sector, leading to surplus for P (2) P has no financial market to save it in (3) when saving it in R R, invests only in simple debt securities. Labor Supply Shock

Globalization and the US Recession

• Why did the recession not occur earlier? ‒ Financial Engineering plus easy money policy response channeled the savings in developing countries into housing in the US ‒ That led to the Housing price bubble ‒ Myopic US households felt wealthy and kept up their consumption ‒ Hid the problems for a while ‒ A As if iin th the stylized t li d model d l we examined i d ‒ The labor employed by R in P saved 0.08 of their 0.18 wages and those savings were invested in R to create a wealth effect to keep consumption in R up

• Some support for this view to follow Labor Supply Shock

China and India – Exports (offshoring?)

China & India: Key Economic Indicators China GDP (US$ billions) Gross domestic investment/GDP (%) Exports of goods & services/GDP (%) Gross domestic savings/GDP (%) India da GDP (US$ billions) Gross domestic investment/GDP (%) Exports of goods & services/GDP (%) Gross domestic savings/GDP g (%)

1982 221.5 33.2 8.9 34.8

1992 454.6 36.2 19.5 37.7

2001 1,167.1 38.5 25.5 40.9

2002 1,232.7 41.0 29.5 44.0

194.8 21.7 6.1 18.3

244.2 23.8 9.0 21.8

478.5 22.3 13.5 23.5

510.2 22.8 15.2 24.2

Labor Supply Shock

Some evidence for this view • Urban population in China increased by nearly 300 million from 1990 to 2007 • A large part of that migration of labor, presumably, was facilitated by offshoring?

Labor Supply Shock

Some evidence: Pressure on US Wages • Wages & Salaries – 2000: 56.7% % of GDP – 2007: 53.7% of GDP – Change, 2000 to 2007: -3.0% of GDP

• However However, as we saw earlier earlier, consumption as a % of GDP increased – Because of housing bubble and perceived increase in wealth – Wall Street Financial Engineering helped create that bubble and also facilitated borrowing against that wealth

• Next: More on Financial Engineering (FE)

Labor Supply Shock

FE facilitated money flow to housing • Origination of non-prime mortgages rose from $500 bn in 2000 to $1500 bn in 2005 • Mortgages Prior to 1990, • Agency mortgage pools ‒ Conforming (<$417K, > 80%LTV,..) ‒ First lien mortgages

• Mortgages During late 90’s 90 s, • Non conforming mortgages ‒ Private label securitization ‒ Asset Backed Securities market boom •

50% market share by 2006

Financial Engineering

Share of Private Label ABS Increased ($/HH)

Financial Engineering

Origination of Non Prime Mortgages Tripled ($Bn)

Financial Engineering

FE led to the Housing Bubble Housing prices increased by 86% S&P/Case – Shiller Home Price Index 2000 Q1: 100.77 2007 Q1: 186.07 Like all bubbles,, housing g bubble also collapsed Wealth effect that kept up consumption vanished Recession followed Financial Engineering

Housing Bubble is Different from Stock Bubble • Stock Bubble collapse p July y 2000/ Sep p 2001 ‒ NASDAQ: Over 60% drop ‒ S&P500: Over 30% drop

• Not followed by as severe a recession ‒ Little wealth effect; Stock holdings of households not leveraged ‒ Easy monetary policy temporarily helped: •

Fed funds: 5.31% in 2001/3 to 2.09% in 2001/11

• Housing Bubble collapse ‒ ‒ ‒ ‒

A 25% drop in home values can wipe out entire home equity Huge wealth effect => severe drop in consumer spending Housing market transactions drop Job mobility severely affected •

Inability to relocate affects speed of recovery

Housing Bubble

Home Ownership More Wide Spread & Levered • Middle three wealth class quintiles, 2004 ‒ Principal residence: 66% of all assets ‒ Corporate equities: 8% of all assets

• Stock holdings ‒ 49% of families ‒ Median value: $24,300

• Primary residence ‒ 68% of families ‒ Median value: $131,000

• Leverage ‒ Mortgage g g debt: 47% of real estate value ‒ Other debt: 7% of other assets value

• Averages understate the severity of leverage ‒ 51% of loans originated in 2006 had CLTV > 80% Source: E.N. Wolff (2007) and A. Arora (2007)

Source: E.N. Wolff, WP, New York University, 2007

Money Channeled into Housing: Bigger Price Effect • Consider a hypothetical yp economy y with 10 households with ‒ $100 in housing per household ‒ $100 in stocks p per household

• Sudden helicopter drop of $10 per household, that they have to bid up prices of stocks or housing • Stocks ‒ ‒ ‒ ‒

Total value of stocks before: $1,000 Total increase in money: $100 %Rise in stock prices: $100/$1 $100/$1,000 000 = 10% Whether everyone invests their $10 or they lend their $10 to others who use it to bid up stock prices does not matter

Housing Bubble

Money Channeled into Housing: Bigger Price Effect … • Housing g ‒ Suppose 9 households give their money to a bank ‒ Bank lends the $90 to one household ‒ That household uses that $90 plus own $10 to bid up the price of one house ‒ % price rise = $100/$100 = 100% ‒ Other households will also think that the value of their homes have increased by 100% (Assessor’s (Assessor s use comparables for home valuation in the US.)

• Leverage more in housing, not as much in stocks ‒ Money multiplier effect is higher in housing market •

Even when individual households can take levered positions in stocks

• Financial Intermediaries that channel money to housing, also are very highly levered, and that accentuates the leverage effects Housing Bubble

Money Channeled into Housing: Bigger Price Effect … • Financial sector that intermediates moneyy flow into housing is also very highly levered

Housing Bubble

Why did the bubble collapse? • Pressure on disposable income of some households triggered defaults ‒ Feed back effect • • • •

Subprime crisis Tightening of credit Larger Financial Crisis Refinancing difficulties

‒ Collapse of housing price bubble

Bubble Collapse

PPI and CPI Changes

Bubble Collapse

PPI and CPI Changes • Pressure on disposable p income of some households triggered defaults ‒ PPI increased sharply from 2004 relative to CPI ‒ Even though g PPI rose, did not g get p passed on to consumers ‒ More reliance on outsourcing ‒ Energy prices rise sharply ‒ Pressure on manufacturing wages and disposable income ‒ Some households default on loans ‒ Bank losses, Flight to safety and Tightening of credit ‒ Real Activities Impaired ‒ Triggers the collapse of the bubble ‒ Great Recession

Bubble Collapse

Tightening of Credit – TED Spread • • • •

Julyy 2006: 32 bp p August 2007: 175 bp October 2008: 457 bp J l 2009 July 2009: 38 bp b

Bubble Collapse

Tightening of Credit: Repo Haircuts

From Arvind Krishnamurthy, “Debt Markets in the Crisis,” 2009, Forthcoming JEPS

Bubble Collapse

Tightening of Credit

Bubble Collapse

How much is the Lost Wealth Effect on Spending? • Residential RE/GDP in 2007: 1.45 • Consumption is 8.4% of change in RE wealth • If RE value drops (permanent) by 25% •

I Impact t on consumption ti ~ (1.45×0.25)×0.084 (1 45 0 25) 0 084 =0.03×GDP 0 03 GDP

• We should expect a permanent drop of about 3% GPD in consumption from 2007 level • We should be able to move on and grow from there

Bubble Collapse

The US is not alone

• These arguments should hold for all countries ‒ That experience a significant CA deficit ‒ Also had a housing price bubble

• CA surplus countries should behave differently

International Evidence

Countries: CA Balance as % of GDP

International Evidence

Countries: Home price inflation relative to CPI

International Evidence

Handling a large labor supply shock & Wealth Destruction: A Lesson From History •

End of World War II – Defense production stopped overnight

•

60 million workers with 16 million returning soldiers – 25% rise in workforce in a very short time

• •

A major part of physical capital of the world destroyed US Department of Labor Forecast – 12 to 15 million unemployed with severe recession

• •

Those predictions did not materialize Policy and Institutional Response – Channeled labor and savings to productive activities – Led to Prosperity instead of Doom

•

Similar response now, Globally

Going Forward

Handling a large labor supply shock & Wealth Destruction: A Lesson From History •

We have to realize that a large part of the wealth we imagined were there are not really there, and move forward

•

Fiscal policy response is needed now as after WWII

•

Development D l t off institutions i tit ti tto channel h l savings i within ithi Chi China, India, … into productive activities

•

Encourage savings in the U U.S.A, S A and not subsidize housing

Going Forward