Private Transfers in Comparative Perspective

Private Transfers in Comparative Perspective Ronald Lee Gretchen Donehower January 9, 2009 National Transfer Accounts Workshop 6 Research support from: NIA R37 AG025247 and R01 AG025488, as well as by grants from MEXT.ACADEMIC FRONTIER (2006-2010) to NUPRI

Outline Introduction Theory of Intergenerational Transfers Evolutionary perspectives Economic theory Measurement of Private Transfers Transfers in the National Transfer Accounts How important relative to GDP? Inter vs Intra household transfers Bequests Age Profiles of Per Capita Transfers Net per capita within household transfers (intrahousehold) Age Profiles of per capita net between household transfers (Interhousehold) Age Patterns of Gross Transfer Flows Within the Household (Intrahousehold) Private Transfers to Elderly Transfers to Children and Investment in Human Capital Private Transfers to Children in Relation to Private Transfers to the Elderly Private vs Public Transfers

In evolutionary past Children were nutritionally dependent until age 20 Massive transfers from parents and others were required to feed them People remained net producers until death Virtually no upward transfers to elderly All transfers were private No asset accumulation

Today Children are still dependent until 20 or later Still require massive transfers But now public sector mediates transfers for education, health care Private transfers are less important for kids Elderly no longer work much Instead, they receive major public transfers Assets are now important and help fund retirement consumption.

What is a transfer? An item of value given to another with no quid pro quo. We can t know if a transaction is a transfer unless we know the motives and expectations of both actors. In NTA we estimate presumptive transfers; we believe they are mostly transfers, but we cannot be sure and must consider whether they are actually exchanges.

Economic Theory Rich theory relating fertility, investments in human capital, which are transfers to children, and old age support. This theory brings in economic context, institutions, and the public sector. Becker Becker and Barro Willis others Not time to present it now; I will sketch it.

Parents are altruistic toward children, but also concerned about own consumption and old age. They make transfers to children for human capital (HK) up to a point. Planned transfers to a child are also the price of that child. The greater the planned transfers, the lower is fertility. Beyond that point, kids may borrow from their parents to get more HK. Repay by supporting parents in old age.

Many contextual factors matter Are the returns to HK greater than interest rate? Are parents rich or poor? Does the culture, religion, or law enforce repayment by children? Is there public education? Are there public pensions?

Theory (cont.) If the rate of return to HK exceeds the market rate of interest, parents limit their fertility in order to invest in each child s HK The optimal investment is to the point where the rate of return drops to the rate of interest. Beyond that, it is better for parents to leave bequests. Parents who are poor or less altruistic want to invest (transfer) less Need money for self

If parents invest less than the optimal amount in each child due to limited altruism or poverty It is efficient for kids to borrow to pay for additional education. But usually can t borrow in market. So perhaps borrow additional money from parents to get more education Later repay loan as old age support. Then transfers to elderly rise as income rises.

Many contextual factors modify the predictions of the theory Are parents highly altruistic toward kids? (culture, religion) Do institutions support repayment of loans to children for higher education (for example)? (culture, society, laws, religion) Are returns to human capital high? Is income low or high? Is there public education? Are there public pensions?

Some predictions Poor countries, low returns to HK High fertility because children have low price Little HK Low familial old age support Perhaps bequests of property. After income and return to HK start to rise More parental investment in HK Moderate fertility because kids have higher price More old age support But what about public sector, pub ed, pub pensions?

Some predictions (cont.) At higher incomes, and high return to HK Heavy investment in HK Very low fertility because kids have high price Parents rich enough to invest optimally in HK and also to provide for own old age So little familial old age support. But at this stage, and previous, much depends on public transfers to children and elderly. Diff outcomes are possible.

Other considerations Are children altruistic toward their parents? Is there heterogeneity across countries in the general strength of altruism in both directions?

Dynastic altruistic utility ( ) ( ) 1 U = V c +δ a n nu j j j j j+

Kinds of private transfers we measure Intrahousehold transfers To children Private education Private healthcare Other consumption (food, housing, clothes, etc.) To elderly (food, housing, clothes, etc.) To spouse To head of household if member has surplus Interhousehold transfers usually money Bequests at death

How do we measure bequests? Surveys give asset holdings by age. Assume mortality is independent of asset holdings We hope to relax this false assumption later Then we have bequests made by age. For each age of death, we can calculate the age distribution of surviving children. We divide up the bequest equally among all of these, so have recipient by age. For childless death, bequest goes to same age sibling. Only a few countries have estimated bequest transfers so far.

How do we measure interhousehold transfers? Category includes Child support following divorce Alimony following divorce Charitable contributions Gifts between family and friends that do not share a household Surveys record payments given and received for households. We assume these all occur to and from the household head These data may be of poor quality, and there are no national control totals to use in adjusting them. Do have aggregate net private transfers at national level These are mostly net remittance flows to and from families abroad. These can be substantial. Some surveys, like the US HRS and European SHARE type surveys have higher quality data on interhousehold transfers for the older population. These will be used to evaluate and improve our current estimates.

How do we measure intrahousehold transfers? Complicated! (Leave explanation, because NTA participants already know all this.)

How do we measure intrahousehold transfers? Measure consumption by each individual (later talk) Measure labor income for each individual

How big are transfers relative to GDP? Calculate aggregate transfers received within and between households (gross flows) Calculate aggregate public transfers that are age-related; ignore non age-related like military, most social infrastructure, other public and quasi public goods. Consider these relative to GDP.

Aggregate transfers relative to GDP for four NTA countries Country Public Age- Related Trans Private Transfers Total Transfers Japan 0.22 0.27 0.49 Taiwan 0.10 0.37 0.47 Thailand 0.06 0.31 0.38 US 0.18 0.26 0.43 Uruguay 0.15 0.51 0.65 Note: In these Does countries, not include private bequests, exceed and public, does and include total share transfers is very between substantial. spouses, for example when only one has labor income.

Do public and private transfers substitute for one another? Early hints. Public age related transfers vs Private transfers (outflows): population weighted sums as shares of GDP Pop wtd sum of public age related trans as share of gdp -0.250-0.200-0.150-0.100-0.050 0.000-0.220 Japan US Taiwan Thailand Pop wtd sum of private trans outflows rel to gdp -0.200-0.240-0.260-0.280-0.300-0.320-0.340-0.360-0.380-0.400 Transfer_shares_of_YL.v3.RDL, tab PubTransVsPrivTrans

Which are bigger, transfers within households or between households? Calculate aggregate transfers received within households and between households. Use same population age distribution to isolate non-demographic differences among countries. Average population age distribution for 23 NTA countries is standard; used later, too.

Ratio of Aggregated Private Interhousehold Transfers Received to Intrahousehold transfers received, weighted by standard population age distribution Ratio of Inter to Intra Priv Trans 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 In the four industrial nations, the Philippines has very high share is always less than.04. ratio, probably due to high foreign remittances. Mexico is next highest, probably for the same reason. 0 Austria 2000 Brazil 1996 Chile 1997 China 2002 Costa Indonesia Rica 2004 2005 Japan 2004 Mexico 2004 Country and Year Philippines Slovenia 1999 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994

The empirical literature on private transfers focuses exclusively on interhousehold transfers or on bequests. The main research question has been whether these transfers reflect altruism or an exchange motive (Cox). We see that this interhousehold transfers are typically only a small fraction of intrahousehold transfers.

Now look at the age distribution of the three kinds of private transfer. First, all three at once, for the US, to give a sense of proportion.

Inter and Intra Household Private Per Capita Net Transfers and Bequests, US, 2003 20000 Within hshld 10000 US $ (2003) 0-10000 0 20 40 60 80 Between hshld tfb tfw tfbb -20000 bequests -30000-40000 Age US.09.26.2008; TF Graphs

Bequests for three countries

0.2 Per Capita Bequests (TFBB) 0 Transfers (Scaled by Avg. YL 30-49) -0.2-0.4-0.6-0.8 Not until age 67 do people in Costa Rica and the US on net begin to make bequests rather than receiving them. Mortality in the two countries is similar. The difference in the bequest schedules reflects differences in the age pattern of fertility and differences in asset holdings relative to labor income. Costa Rica 2004 Indonesia 2005 US 2003-1 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 Age

Private net intrahousehold transfers, for 14 countries. Standardized by dividing through by labor income in each country, averaged over ages 30-49. Units for transfers are years of labor income. Average labor income is very highly correlated with per capita GDP.

0.8 Per Capita Private Intrahousehold Transfers (TFW) 0.6 Transfers (Scaled by Avg. YL 30-49) 0.4 0.2 0-0.2-0.4-0.6-0.8-1 -1.2 1 In many countries elder household members continue to make transfers to younger ones until extreme old age. Includes Brazil, Philippines, Mexico, Uruguay, Indonesia, and all Western. 5 9 13 17 21 25 29 33 37 41 45 49 Age 53 57 61 Only Taiwan, and to a lesser degree Thailand and China, show expected pattern of elder co-residence. 65 69 73 77 81 85 89 Austria 2000 Brazil 1996 Chile 1997 China 2002 Costa Rica 2004 Indonesia 2005 Japan 2004 Mexico 2004 Philippines 1999 Slovenia 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994

Why do co-resident elders continue to net transfer to others in household? Remittance income will be attributed to them if they are household head, and then transferred by them to others. (Phil., Mex) Some continue to have quite high labor income. (Indonesia, Philippines) Some may receive generous pension benefits (Brazil). Some may have accumulated assets, e.g. own the farm, so property income comes to them if they are heads.

0.8 Per Capita Private Intrahousehold Transfers (TFW) 0.6 Transfers (Scaled by Avg. YL 30-49) 0.4 0.2 0-0.2-0.4-0.6-0.8-1 -1.2 1 5 Some av prime age workers transfer up to 80 or 90% of average labor income. 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 In Western nations, these prime age workers transfer only about 20 to 25% of labor income. 69 73 77 81 85 89 Austria 2000 Brazil 1996 Chile 1997 China 2002 Costa Rica 2004 Indonesia 2005 Japan 2004 Mexico 2004 Philippines 1999 Slovenia 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994 Age

Transfers (Scaled by Avg. YL 30-49) 0.8 0.6 0.4 0.2 0-0.2-0.4-0.6-0.8-1 -1.2 1 5 9 13 Per Capita Private Intrahousehold Transfers (TFW) Western industrial nations make low transfers to kids, probably because of public education and health care. China also stands out for low transfers to children. 17 21 25 29 Teenagers in Brazil and Uruguay get net transfers of 60% of labor inc, perhaps for private education. Taiwan, Mexico, Chile and Philippines are close. Probably for private education. 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 Austria 2000 Brazil 1996 Chile 1997 China 2002 Costa Rica 2004 Indonesia 2005 Japan 2004 Mexico 2004 Philippines 1999 Slovenia 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994 Age

Transfers (Scaled by Avg. YL 30-49) 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0-0.05-0.1 1 5 Per Capita Private Interhousehold Transfers (TFB) Philippines have largest by far, reflecting remittance income received. Mexico and Thailand are also high. Transfers received is highest in old age. The US has negative interhh transfers all through adulthood, perhaps due to remittances sent back to Mexico, Philippines, and China. Similarly, Austria and Slovenia. 9 13 17 21 25 29 33 37 41 45 49 Age 53 57 61 65 69 73 77 81 85 89 Austria 2000 Chile 1997 China 2002 Costa Rica 2004 Indonesia 2005 Japan 2004 Mexico 2004 Philippines 1999 Slovenia 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994

Transfers (Scaled by Avg. YL 30-49) 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0-0.05-0.1 1 5 Per Capita Private Interhousehold Transfers (TFB) Taiwan again shows the expected classic effect of making transfers in working years and receiving them when old, the support pattern for nonco-resident elderly. 9 13 17 21 25 29 33 37 41 45 49 Age 53 57 61 65 69 73 77 81 85 89 Austria 2000 Chile 1997 China 2002 Costa Rica 2004 Indonesia 2005 Japan 2004 Mexico 2004 Philippines 1999 Slovenia 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994

Now look at age pattern of gross transfers within the household for US and Japan.

0 The strong diagonal reflects transfers between spouses. Perhaps these should not be counted. Same age. Division of labor. 5,500 Intrahousehold Transfers in the US (Aggregate in Millions) Shows aggregate amount of transfers between two ages. Not per capita. Affected by pop age distribution. 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500-45 60 75 90 5,000-5,500 4,500-5,000 4,000-4,500 3,500-4,000 3,000-3,500 2,500-3,000 2,000-2,500 1,500-2,000 1,000-1,500 500-1,000 - - 500 10 20 30 40 50 15 30 Transfers Made (From Age) Transfers Received (To Age) 60 70 80 90 0

0 Transfers to 0-4 from parents age 25-29. Intrahousehold Transfers in the US (Aggregate in Millions) Transfers 5,500 Transfers to college age kids from parents age 40-49. 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 - Big peak at old age is spousal transfers, due to: a) one is head and is assumed to own asset, e.g. home; b) one may have pension 90 that is 75 shared. 60 45 5,000-5,500 4,500-5,000 4,000-4,500 3,500-4,000 3,000-3,500 2,500-3,000 2,000-2,500 1,500-2,000 1,000-1,500 500-1,000 - - 500 Note that children age 10-14 and 15-19 make significant transfers. 10 20 30 40 50 60 Transfers Received (To Age) 70 80 90 0 15 30 Transfers Made (From Age)

0 Higher education paid by older adults. Intrahousehold Transfers in Japan (Aggregate in Millions of Yen) 7,000 6,500 6,000 5,500 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 - Japan is tidier. Children don t contribute till 25 or so. 10 20 30 40 Later age at childbearing also changes appearance. 50 Transfers Received (To Age) 60 70 Spousal diagonal 80 90 30 15 0 90 This little hump is 75 familial 60 support of the 45 elderly. Remarkably small. Transfers Made (From Age) 6,500-7,000 6,000-6,500 5,500-6,000 5,000-5,500 4,500-5,000 4,000-4,500 3,500-4,000 3,000-3,500 2,500-3,000 2,000-2,500 1,500-2,000 1,000-1,500 500-1,000 - - 500

Now look more closely at private transfers to the elderly. Summarize by summing the net transfers over ages 60-79, divided by av labor inc. Plot against per capita GDP, PPP adjusted.

Private Transfers per Elder (Sum 60-79/yl(30-49) Vs Per capita GDP, by Region Priv Trans to Old (sum 60-79)/av yl(30-49) 6 4 2 0-2 -4-6 -8-10 China Philippines Costa Rica Indonesia Thailand Brazil Mexico Chile Uruguay Taiwan Slovenia Austria Japan 0 5000 10000 15000 20000 25000 30000 35000 40000 In some countries, the elderly make large net transfers to others: Brazil, Indonesia, Uruguay, and Mexico. GDP per capita, PPP adj The elderly in rich industrial nations make small net transfers to others, including Japan. Blue = Asia Red = Latin America Green = Western Industrial GDP_PPP US

Private Transfers per Elder (Sum 60-79/yl(30-49) Vs Per capita GDP, by Region Priv Trans to Old (sum 60-79)/av yl(30-49) 6 4 2 0-2 -4-6 -8-10 China Philippines Costa Rica Indonesia Thailand Brazil Mexico Chile Uruguay Taiwan In only a few countries, China, Thailand, and Taiwan, do the elderly receive positive transfers. Slovenia Austria Japan 0 5000 10000 15000 20000 25000 30000 35000 40000 GDP per capita, PPP adj Blue = Asia Red = Latin America Green = Western Industrial GDP_PPP US

Priv Trans to Old (sum 60-79)/av yl(30-49) 6 4 2 0-2 -4-6 -8-10 Private Transfers per Elder (Sum 60-79/yl(30-49) Vs Per capita GDP, by Region The Asian countries show a pattern of transfers to the elderly increasing Taiwan with per capita GDP, and then falling at the highest level, perhaps Thailand due to generous public pensions, China health care, and long term care. Philippines Costa Rica Indonesia Brazil Mexico Chile Uruguay Slovenia Austria Japan 0 5000 10000 15000 20000 25000 30000 35000 40000 GDP per capita, PPP adj Blue = Asia Red = Latin America Green = Western Industrial GDP_PPP US

Priv Trans to Old (sum 60-79)/av yl(30-49) 6 4 2 0-2 -4-6 -8-10 Private Transfers per Elder (Sum 60-79/yl(30-49) Vs Per capita GDP, by Region China Philippines Costa Rica Indonesia Thailand Brazil Mexico Chile Uruguay Taiwan Slovenia Austria Japan 0 5000 10000 15000 20000 25000 30000 35000 40000 GDP per capita, PPP adj The LA countries do not show so clear a picture. At any level of income, the Latin American countries make smaller private transfers to elderly than do Asians. Blue = Asia Red = Latin America Green = Western Industrial GDP_PPP US

Living Arrangements and Intrahousehold Inflows to, and Outflows from, the Elderly Avg Interhh Inflow 60+/ Avg YL 30-49 0.450 0.400 0.350 0.300 0.250 0.200 0.150 0.100 0.050 0.000 Living Arrangements vs IntraHH Inflows TH PHMX CR ID BR CL TW JP y = -0.003x + 0.361 R 2 = 0.3687 0.0 20.0 40.0 60.0 80.0 AT Prop. Age 60+ Living Alone or in Couple US Avg Inrahh Outflow 60+/ Avg YL 30-49 0.000-0.100-0.200-0.300-0.400-0.500-0.600-0.700-0.800 Living Arrangements vs IntraHH Outflows TH CR MX PH CL ID BR TW JP AT y = 0.0059x - 0.5755 R 2 = 0.3211 0.0 20.0 40.0 60.0 80.0 Prop. Age 60+ Living Alone or in Couple US When more elderly live alone or only with spouse, the average elder receives smaller within household transfers, and makes smaller within household transfers.

Completely obvious result And similarly, interhousehold flows must increase when more elderly live alone. Do they?

Living Arrangements vs InterHH Inflows 0.400 Avg Interhh Inflows to 60+/ Avg YL 30-49 0.350 0.300 0.250 0.200 0.150 0.100 0.050 0.000 TH PH MX CR ID CL BR TW No, they also decrease when more elderly live alone. JP AT y = -0.0028x + 0.1834 R 2 = 0.2573 Because elderly live alone when they can afford to, due to public pensions or assets. Endogenous. Don t need transfers. 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 US -0.050 Prop. Age 60+ Living Alone or in Couple

Now look more closely at private transfers to children. Summarize by summing the net transfers over ages 0 to 19, divided by av labor inc.

Priv Trans per child (sum 0-20)/yl(30-49) 14 12 10 8 6 4 2 0 Brazil Costa Rica Private Transfers per Child vs GDP per capita Uruguay Mexico Chile Indonesia Philippines Thailand China All receive positive private net transfers, from 5 to 13 years of per capita income. Taiwan Slovenia Japan Austria 0 5000 10000 15000 20000 25000 30000 35000 40000 GDP per capita, PPP adj The Asian countries again show a rough inverted U pattern, consistent with the theory. The theory says that an increasing share of transfers to kids are actually loans, subject to repayment in old age. Consistent with elder inverted U pattern. GDP_PPP US The rich industrial nations have low transfers to kids, perhaps due to strong public education. China also has low transfers to kids.

Private Transfers per Child vs GDP per capita Priv Trans per child (sum 0-20)/yl(30-49) 14 12 10 8 6 4 2 Brazil Costa Rica Uruguay Mexico Chile Indonesia Philippines Thailand China For me, a big surprise to see Latin American countries transferring so heavily to kids, much more than the Asian countries. Taiwan Slovenia Japan Austria US 0 0 5000 10000 15000 20000 25000 30000 35000 40000 GDP per capita, PPP adj GDP_PPP

What share of HK investment in kids is privately financed? Construct a measure of HK investment per child by summing all public and private spending on health care and on education from age 0 to 18 for health, to 26 for educ. Divide by av labor income, as usual. Calculate the share of the total that is private spending.

Share of Private spending in HK investment up to age 18 or 26 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 Private share of Human Capital investment per child, by Per Capita GDP, 23 NTA countries Great variation, from China at China 72% to Sweden at 3%. All the industrial nations are low, and most are <10%. Kenya Philippines India Indonesia Mexico Brazil Uruguay Thailand Chile Taiwan South Korea Costa Rica Spain Hungary Japan Slovenia Germany Austria Finland France Sweden US 0.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50 10.00 10.50 11.00 GDP Per Capita, PPP Adj

The Quantity-Quality Tradeoff Economic theories emphasize the tradeoff between numbers of children, and investment per child. Here look at HK spending (as defined above) and the current Total Fertility Rate.

Figure 13. Per Child HK Spending (Public and Private) vs. Fertility ln(hk per Child/Av Lab Inc 30-49) 2.00 1.50 1.00 0.50 0.00 SI TWJP ES HUAT DE KR CN SE FR FI TH US MX CR CL ID A strong negative relationship, with a slope close to -1 (loglog). BR UY IN PH y = -0.928x + 1.8163 R 2 = 0.5984 Amount spent on HK is around 5 or 6 years of labor income, regardless of fertility. KE -0.50 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 ln(tfr)

Ratio of private transfers received by old to young, using standard population weights 0.25 Priv Trans Rcvd by Young/Old 0.2 0.15 0.1 0.05 0 Austria 2000 Brazil 1996 Chile 1997 China 2002 Costa Indonesia Rica 2004 2005 Japan 2004 Mexico 2004 PhilippinesSlovenia 1999 2004 Taiwan 1998 Thailand 2004 US 2003 Uruguay 1994 Country and Year TF_11.21.08RDL.v3

Ratio of Private Transfers per Elder to Youth Vs Per capita GDP, by Region Ratio of Priv Trans to Old vs Young 0.6 0.4 0.2 0-0.2-0.4-0.6 ChinaThailand Philippines Costa Rica Indonesia Brazil Chile Mexico Uruguay Taiwan Slovenia Japan 0 5000 10000 15000 20000 25000 30000 35000 40000 Austria Blue = Asia Red = Latin America Green = Western Industrial US -0.8 GDP per capita, PPP adj GDP_PPP

Private transfers received per elder vs private transfers received per child Pub Trans Rcvd by Elder, sum 60-79, rel to yl(30-49) 6 4 2 0-2 -4-6 -8 y = -0.738x + 4.7224 R 2 = 0.323 China Thailand Austria Slovenia US Japan Costa Rica Philippines Chile Indonesia Taiwan 0 2 4 6 8 10 12 14 Mexico Uruguay Brazil -10 Priv Trans Rcvd by Child, summed 0-24, rel to yl(30-49)

Private transfers received per elder vs private transfers received per child Pub Trans Rcvd by Elder, sum 60-79, rel to yl(30-49) 6 4 2 0-2 -4-6 -8 China Thailand Austria Slovenia US Japan Costa Rica Philippines Chile Indonesia Taiwan 0 2 4 6 8 10 12 14 Countries with higher transfers to children do not necessarily have higher transfers to the elderly. More the opposite. Mexico Uruguay Brazil -10 Priv Trans Rcvd by Child, summed 0-24, rel to yl(30-49)

Transfers to Elder vs HK Spending per child 3.5 HK Spending per child (sum 0-18 or 26) Brazil Uruguay Indonesia Mexico Japan 1.5 Philippines ChileUS Costa Rica 3 2.5 2 1 Slovenia 0.5 Austria China Thailand Taiwan 0-10 -8-6 -4-2 0 2 4 6 Transfers per Elder (sum 60-79)

END