Household Income, Demand, and Saving:

Deriving Macro Data with Micro Data Concepts

Barry Cynamon

Frontiers of Measuring Household Economic Behavior

Federal Reserve Bank of Boston, April 27, 2015

Multi-year research project linking household finances and economic growth

Acknowledgements

Multi-year research project linking household finances and economic growth

Acknowledgements

• Joint work with Steve Fazzari • Generous support from INET

• Opinions are mine and not those of the Fed

Reconciling macro and micro estimates of U.S. household income and expenditures

This Session

Reconciling macro and micro estimates of U.S. household income and expenditures

This Session

• Understand how the aggregate measures are distributed

• Validate survey measures comparing to trusted aggregate measures

• Learn from aggregates consistent with micro data concepts

INEQUALITY AND CONSUMPTION Motivation for Measurement

“Inequality, the Great Recession, and Slow Recovery” Forthcoming in the Cambridge Journal of Economics Working paper available at SSRN: http://ssrn.com/abstract=2205524

Investigate relationship between income inequality and Great Recession

The Original Goal

Investigate relationship between income inequality and Great Recession

The Original Goal

• Rich have lower propensity to consume

– (Maki and Palumbo 2001)

• Increasing share of income flowing toward rich

– (Piketty and Saez, 2003; CBO, 2013; Johnson and Smeeding, 2014)

• Downward pressure on aggregate consumption ?

Increasing share of income flowing toward rich

Income share of the top 5% of US households

Source: The World Top Incomes Database

15%

20%

25%

30%

35%

40%

1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

Consumption Drove Economic Growth

Consumption share of GDP

55%

60%

65%

70%

1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

Source: BEA National Income and Product Accounts

Households Doubled their Leverage Debt to income ratio of US households

Source: FRB Financial Accounts of the United States

40%

60%

80%

100%

120%

140%

1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

Find a micro data set with income and consumption expenditure

Initial Plan

Find a micro data set with income and consumption expenditure

Initial Plan

• SCF: oversamples wealthy, but no consumption data • CPS: annual and large sample, but no consumption

• CE: fails to match aggregate data in level or trend

– Under-reporting especially among higher income households (Sabelhaus, Johnson, Ash, Swanson, Garner, Greenlees, Henderson, 2013)

Use a mix of aggregate and micro data to generate results at “group” level

Revised Plan

Revised Plan

• SCF: for distribution of balance sheet accounts

• CPS: for distribution of income

• National accounts: for authoritative time series

Use a mix of aggregate and micro data to generate results at “group” level

Maki and Palumbo (2001)

• Numbers add up to net worth and saving for the

personal sector published in the FAOTUS

• Distribution matches the SCF in every survey year

Assets and Liabilities Income

Stocks Flows

Aggregate FAOTUS FAOTUS NIPA

disposable personal income

Micro SCF *identification CPS

shares interpolated linearly between waves

assume flows proportional to holdings

money income

M&P for group-level saving numbers and then back out consumption numbers

Revised Plan ii

M&P for group-level saving numbers and then back out consumption numbers

Revised Plan ii

• Mark Zandi provided us with saving rate information

derived using the M&P approach

• First, we adjusted those FAOTUS saving numbers to match NIPA saving numbers

• Then we allocated NIPA transfers and interest between our groups so we could back out “group” consumption

M&P for group-level saving numbers and then back out consumption numbers

Revised Plan ii

Disposable Income

Saving Outlays − =

Consumption Transfers Interest Outlays = + +

Consumption = Disposable Income

− Saving − − Transfers Interest

Non-rich took on debt to maintain consumption, which delayed the effect of rising inequality

The Story

Non-rich took on debt to maintain consumption, which delayed the effect of rising inequality

The Story

Before Great Recession:

• Debt to income ratio of non-rich grew before GR

• Consumption rate of non-rich stable or rising After Great Recession:

• Consumption of rich only has recovered

• Per capita, real GDP far below trend after Great Recession

Non-rich took on debt to maintain consumption, which delayed the effect of rising inequality

The Story

Before Great Recession:

• Debt to income ratio of non-rich grew before GR

• Consumption rate of non-rich stable or rising After Great Recession:

• Consumption of rich only has recovered

• Per capita, real GDP far below trend after Great Recession

Debt to income ratio of non-rich increased

25%

50%

75%

100%

125%

150%

175%

200%

Bottom 95%

Top 5%

Source: FRB Survey of Consumer Finances, data provided by Romain Ranciere

Consumption rate of non-rich stable or rising

70%

75%

80%

85%

90%

95%

100%1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Consumption Rate 95% Consumption Rate 5%

Outlay Rate 95% Outlay Rate 5%

Source: Cynamon and Fazzari (2015)

Consumption of rich has recovered; that of non-rich has not

80

100

120

140

160

180

200

220

240

260

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

Index of Real Consumption, Bottom 95% and Top 5% (1989=100)

Bottom 95% Top 5%

Source: Data from Cynamon and Fazzari (2015)

GDP well below trend after Great Recession

Per capita, real GDP, chained dollars (exponential trend)

Source: BEA National Product Accounts

$5,000

$20,000

$35,000

$50,000

$65,000

1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

MEASURING DEMAND Measurement Exercise

“Household Income, Demand, and Saving: Deriving Macro Data with Micro Data Concepts” Forthcoming in the Review of Income and Wealth Working paper available at SSRN: http://ssrn.com/abstract=2211896

Motivation

Reconcile macro and micro estimates of U.S. household income and expenditures

Motivation i

• Maki and Palumbo (2001) reliant on consistent concepts – CPS income distribution applied to NIPA disposable personal income – SCF net worth distribution applied to FAOTUS balance sheet

• But there are inconsistencies between micro and macro data – Not just sampling error; important conceptual differences

• Previous efforts to match NIPA and survey income – Katz (2012), Bosworth et al. (2007)

Reconcile macro and micro estimates of U.S. household income and expenditures

Comparisons to Survey Data

CPS SCF

PSID

0.60

0.70

0.80

0.90

1.00

1.10

1988

1991

1994

1997

2000

2003

2006

2009

2012

0.60

0.70

0.80

0.90

1.00

1.10

1967

1970

1973

1976

1979

1982

1985

1988

1991

1994

1997

2000

2003

2006

2009

2012

0.60

0.70

0.80

0.90

1.00

1.10

1969

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

Pre-tax income data from surveys well below 100% of

NIPA personal income

Motivation ii

• Might learn from macro measures adjusted to match micro concepts – PCE vs. what households actually spend

– Different definitions of saving may tell different stories

Reconcile macro and micro estimates of U.S. household income and expenditures

Objective

• Eliminate imputed value of services in consumption – Example: Imputed rent

• Eliminate spending not controlled by households – Example: Medicare

Measure the flows of purchasing power under the control of the household

Objective

• Household financial flows the way households actually see these flows

• Concept likely to correspond better with flows households report on surveys

Measure the flows of purchasing power under the control of the household

Key Identity

• Accounting identity maintained before and after adjustments:

Disposable Household Transfers Financial Income = Consumption + Investment + & Interest + Saving

• Identity holds in NIPA – Household investment not distinguished from financial saving

• Adjustments to consumption or income require balancing change elsewhere

Housing Example (2013 $billions)

Disp. Income

Cons. HH Invest. Trans. & Int.

Fin. Saving

Implicit Rent - 1326 - 1326

Intermediate Inputs + 152 + 152

Mortgage Interest + 334 + 334

Depreciation + 312 + 312

New Construction Single-Family Homes

+ 426 - 426

Broker commissions + 105 - 105

Total - 528 - 1068 + 321 + 334 - 115

• Eliminate “rent home to yourself” business

Other Important Adjustments

• About 40 separate adjustments

• Remove non profit institutions that serve households

• Free financial services

• Medical care – Employer and government, not households

• Retirement accounting – Exclude contributions by employers and government to defined

benefit plans – Include benefits from DB plans

Other Important Adjustments

Category Disposable

Income Consumption

Transfers & Interest

Financial Saving

Owner-Occupied Housing

-4% -9% 81% -19%

Financial Services -6% -2% -76%

Defined Benefit Pensions

-1% -27%

Third-Party Paid Medical Services

-13% -14%

Non-Profit Sector -1% -4% 61% 8%

Other -2% -30%

Adjusted Data 73% 70% 242% -44%

Note: Household investment excluded form table, because it has no clear personal sector counterpart in the NIPA

Adjusted measures: real, per capita

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000

$35,000

$40,000

19481953195819631968197319781983198819931998200320082013

Disposable Personal Income

Adjusted Disposable Income

$0

$1,000

$2,000

$3,000

$4,000

$5,000

19481953195819631968197319781983198819931998200320082013

Personal Interest and Transfers

Adjusted Transfers and Interest

$0

$5,000

$10,000

$15,000

$20,000

$25,000

$30,000

$35,000

$40,000

19481953195819631968197319781983198819931998200320082013

Personal Consumption ExpendituresHousehold DemandAdjusted Consumption

-$4,000

-$3,000

-$2,000

-$1,000

$0

$1,000

$2,000

$3,000

$4,000

19481953195819631968197319781983198819931998200320082013

Personal SavingAdjusted Gross Household SavingFinancial Saving

Disposable Income Transfers and Interest

Consumption Saving

Comparisons to Survey Data

CPS SCF

PSID

0.60

0.70

0.80

0.90

1.00

1.10

1988

1991

1994

1997

2000

2003

2006

2009

2012

0.60

0.70

0.80

0.90

1.00

1.10

1967

1970

1973

1976

1979

1982

1985

1988

1991

1994

1997

2000

2003

2006

2009

2012

0.60

0.70

0.80

0.90

1.00

1.10

1969

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

Note: All measures shown pre-tax; CBO net realized capital gains added to adjusted disposable income to match SCF, which includes realized gains.

Expenditure Shares of Income

70%

75%

80%

85%

90%

95%

100%

105%

110%

115%

1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013

Adjusted Consumption Household Investment Adjusted Transfers and Interest

0%

Bigger Collapse: Cash Flow Measure

82%

84%

86%

88%

90%

92%

94%

96%

98%

100%

102%

19

48

19

50

19

52

19

54

19

56

19

58

19

60

19

62

19

64

19

66

19

68

19

70

19

72

19

74

19

76

19

78

19

80

19

82

19

84

19

86

19

88

19

90

19

92

19

94

19

96

19

98

20

00

20

02

20

04

20

06

20

08

20

10

20

12

Demand Rates: NIPA Definition and Adjusted

NIPA PCE / NIPA DPI Adj HH Dem / Adj DPI

New saving rate concepts

-15%

-10%

-5%

0%

5%

10%

15%

1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013

NIPA Saving Rate Adj. Gross Household Saving Rate Adj. Financial Saving Rate

Future Directions

1. Use the Maki and Palumbo procedure with micro-consistent aggregate income and saving series; see if the results change – Wondering if anybody at the Board would like to team up with us

2. Investigate the business cycle properties of the micro-consistent aggregate consumption series – Would like to generate quarterly-frequency numbers

3. Exploit panel structure of PSID to see if story of rising balance sheet fragility among non-rich followed by discrete fall in consumption during GR holds up at household level – Joint work with Daniel Cooper

Future Directions

1. Use the Maki and Palumbo procedure with micro-consistent aggregate income and saving series; see if the results change – Wondering if anybody at the Board would like to team up with us

2. Investigate the business cycle properties of the micro-consistent aggregate consumption series – Would like to generate quarterly-frequency numbers

3. Exploit panel structure of PSID to see if story of rising balance sheet fragility among non-rich followed by discrete fall in consumption during GR holds up at household level – Joint work with Daniel Cooper

Future Directions

1. Use the Maki and Palumbo procedure with micro-consistent aggregate income and saving series; see if the results change

– Wondering if anybody at the Board would like to team up with us

2. Investigate the business cycle properties of the micro-consistent aggregate consumption series – Would like to generate quarterly-frequency numbers

3. Exploit panel structure of PSID to see if story of rising balance sheet fragility among non-rich followed by discrete fall in consumption during GR holds up at household level – Joint work with Daniel Cooper