Für neue Autoren:
kostenlos, einfach und schnell
Für bereits registrierte Autoren
60 Seiten, Note: 1,7
Tables and figures
2. Household saving
2.1 Measuring household saving
3. The decline in U.S. household saving
4. Explaining thesavings drop
4.2 External factors: governmental policis and market innovation
4.3 Income and wealth: the paradox of growth and declining saving rates
4.3.2 Implications: the wealth effect
5. Creating a micro-level sample to measure saving
5.10verview of the PSID
5.1.1 PSID research ares
5.2 Active and passive saving
5.3 Variables and sample selection
5.3.1 Wealth variables
5.3.2 Flow variables
5.3.3 Income variables
5.3.4 Sample selection and general thoughts
6.1 Full sample
6.2 Income groups
6.3 Portfolio composition and net investment
Table 1 : PSID wealth components
Table 2: Mean saving rates, full sample, 1984-2007
Table 3: Summary statistics of active and passive saving, 1984-2007
Figure 1: NIPA and FFA U.S. household saving rate
Figure 2: U.S. household saving rates in percent, 1900-2005
Figure 3: U.S. household sector growth in incomes and wealth
Figure 4: U.S. household saving rate and wealth-to-income ratio, 1960-2000
Figure 5: Income inequality in the U.S
Figure 6: Inequality in the U.S.: realized capital gains, 1980-2003
Figure 7: Inequality and savings drop
Figure 8: PSID sample schematic
Figure 9: Isolating active from passive saving
Figure 10: Substitution of active saving with passive saving
Figure 11 : The composition of saving across different income groups, 1984-2007
Figure 12: Active and passive saving, 1984-2007
Figure 13: Mean dollar amount of annual active saving across income groups
Figure 14: Average annual active saving across income groups, 1984-200
Figure 15: Average annual wealth growth across income groups, 1984-2007
The U.S. household saving rate has dropped remarkably over the past three decades, falling from about 10 percent of disposable income in the early 1980s to near-zero values at the turn of the millennium. After decades of overall stable personal savings following the end of World War II, recently the household savings rate has reached a new all-time low after the Great Depression. However, the causes for this unprecedented decline in savings byAmerican households remain a mystery. Numerous researchers have offered different explanations of the savings drop, including demographic changes, market growth and the rising proportion of capital gains, income and wealth inequality, governmental policies, and innovations in finance and credit.
However, regarding the studies done so far, it is to say that no consensus concerning the underlying reasons of the saving reduction is achieved at all. While scholars have analyzed the many different aspects of the savings drop, an agreement on the fundamental causes is still out of reach. The falling saving rate within the past 25 years of economic expansion is unique in history: while former patterns of household saving could be explained by severe political and macroeconomic changes, the recent substantial drop in times of economic prosperity seems more difficult to investigate. Comparable developments are hardly to be found in U.S. history as well as in other countries’ statistics.
Moreover, mainstream economic concepts and a major part of evidence suggest a positive correlation between income, wealth and saving. Hence, it seems striking that the decline in savings took place in a period of overall economic growth, where incomes and in particular the wealth-to-income has been growing.
The purpose of this paper is to estimate how different income and wealth levels have been related to the household saving patterns within the last three decades in order to disentangle a further piece of the savings drop-puzzle. How was saving
distributed across different income groups? How does growing income and wealth inequality relate to the distribution of savings and what is the actual impact of rising capital gains on saving patterns?
This paper makes use of micro-level household data in order to fully capture the evolution of the saving and wealth of American households over the past three decades. The patterns of personal saving in between the years of 1984 and 2007 will be examined using a comprehensive household data set provided by the Panel Study of Income Dynamics (PSID) by the Survey Research Center at the University of Michigan. I will apply the unique PSID measure of personal savings, which allows separating effective saving efforts from passive capital gains on assets.
First, I discovered that household saving has dropped in all income groups. Secondly, I observed that the decline of the saving-to-income ratio was more severe among middle and low-income households than in the top quintile of the distribution. However, the upper-income group may have accounted for a portion of the overall decline due to unequally higher total wealth and saving levels. Third, my results show that passive saving in the form of unrealized capital gains on assets has merely displaced effective saving measures over the past decades. Hence, possibly rising equity prices were a growing incentive to prefer consumption over saving. Generally, I conclude that the changing patterns of income and wealth can explain a portion of the change in household saving.
In the next section, the measurement of household saving as well as mainstream theoretical concepts and empirical results on personal saving behavior will be presented. Section 3 provides an overview of the recent savings drop in historical perspective. In section 4 potential explanations for the decline in household saving are discussed. First, explanations concerning demographics and external factors such as government policies and financial innovations are reviewed. Secondly, the theoretical implications of the paradox of falling saving rates in times of (unequal) income and wealth growth are presented. In section 5 I develop a measure of household saving using PSID data. After a general overview of the PSID, I explain the methodology applied to separate active and passive saving. Following, the sample selection and the treatment of the key variables is presented. Section 6 provides the detailed summary statistics. Finally, my general conclusions as well as further thoughts on the findings are presented in Section 7.
Generally, saving is referred to as current resources or production that is not consumed in the current period but rather made disposable for future consumption. Following this approach, saving can be broadly defined as income less consumption or, alternatively, the change in one’s wealth. Given a consistent and comprehensive definition of consumption, income, and wealth, both concepts would produce the same empirical results of saving.
In reality, however, even if a single approach of the two is accepted, saving measures can vary in concept and scope. Diverging goals of analysis and levels of aggregation lead to different definitions of what is included and excluded in saving measures. In return, existing predefinitions ofwealth, income, and consumption determine the scope and concept of measurement.
In this paper, a narrow measure of saving is examined, namely the household (or personal) saving, most commonly expressed as a ratio to disposable (after-tax) current-period household income.
In the United States, two different household saving measures exist at the aggregate or national level: personal saving within the National Income and Product Account (NIPA) and the Flow of Funds Account (FFA). As suggested above, the two approaches do not produce the same results as the general methodology and the treatment ofwealth, consumption, and income components differs.
Under NIPA, private saving is a residual, measured on a monthly basis. It equals the difference between personal consumption outlays and personal disposable income. Monthly personal saving divided by disposable income makes up the personal saving rate. In contrast, under FFA, personal saving is measured directly as the household sector’s net acquisition of assets less its net accumulation of liabilities on an annual basis. Hence, the FFA annual personal saving rate equals the net change in household wealth divided by disposable income. Further, the methodology differs regarding certain income and wealth components. FFA treats the net purchase of long-term consumer durables (products that can be used for several years) as saving whereas NIPA considers such expenditures as personal consumption. Moreover, the consideration of capital gains differs between NIPA and FFA. FFA includes realized capital gains in personal income, whereas the NIPA does not.
It is worth mentioning that both approaches ignore unrealized capital gains on assets. In this context, it is to say that the official measures are criticized for their high level of aggregation, for not being transparent, nor being fully representing household-level economic concepts of saving. Researchers interested in saving patterns across certain groups of the population can gather no sufficient information from the aggregate numbers of FFA and NIPA.
There exists a variety of household data sets in the U.S. from which information on the development of saving on the micro-level can be derived. These sources cover saving, and also wealth, of several population groups and over different periods. The three most comprehensive, longest running and representative datasets are the Survey of Consumer Finances (SCF), the Consumer Expenditures Survey (CES), and the Panel Study of Income Dynamics (PSID). As with NIPA and FFA, household saving and wealth measures diverge: CES calculates saving directly as income minus consumption whereas SCF and PSID approach saving as the difference in wealth. Moreover, PSID accounts for unrealized capital gains by separately identifying active and passive saving components.
Before further discussing the impact of different measures and the evidence from the aggregate and household level data sources on the savings drop, let us start by surveying ‘the state of the art’ of theory and evidence on household saving motives and determinants.
Ever since John Maynard Keynes (1936) influential basics, economists used both theory and empirics to assess the motives and determinants of saving. In the following part, I will shortly review the major contributions to the analysis of saving and consumption patterns.
In a classical Keynesian model, referred to as the absolute income hypothesis, saving is a positive function of current-period income. That is households tend to not spending all of their income, but to save more as income grows due to a decreasing marginal propensity to consume. Although Keynes already anticipated a remarkable number of powerful motives to save from which many gained attention later, his theory was soon to be modified and advanced.
The second early theory is the relative income hypothesis. James Duesenberry (1949) observed that the percentage of income saved or consumed would not be solely determined by the individual income level, but rather by income in relation to other consumers. Thus, the percentage position in the income distribution is crucial to consumption and saving levels. Moreover, he figured that current consumption and saving may depend on income levels of past periods. That is, consumption is determined by the highest level of income reached so far instead of current income. Despite his success in incorporating time and group effects, Duesenberry’s explanation of saving behavior was quickly replaced by a model that is well-known and also heavily debated until today: the lifecycle hypothesis.
In the early 1950s, in attempting to explain the difference in aggregate crosscountry saving patterns, Franco Modigliani presented evidence for the fact that saving and consumption are basically uncorrelated to current-period income levels, but in fact determined by long-term (or lifetime-) future income. That is, expectations are crucial as saving and consumption would be smoothed out towards retirement over a lifetime in order to ensure a certain standard of living.
Hence, in the so-called lifecycle model of saving, aging and pension saving play an important role: younger people would consume and may even borrow more as they expect their incomes to grow. Middle-age people save more to pay back past borrowing and to accumulate pension wealth. When people leave workforce, they tend to spend their accumulated retirement savings without having to fear potential future losses. This would lead to a hump shaped saving distribution by age.
Based on Modigliani’s findings, Milton Friedman raised the complementary permanent-income hypothesis which also assumes that people mainly discount shortterm fluctuations in income when determining current levels of consumption and saving. He suggested that one would save more if the current income is higher than the expected long-term income and less when income is expected to grow. To sum up, under the permanent-income lifecycle theories, in contrast to the absolute and relative-income hypothesis, long-term income determines current saving whereas shortterm increases and decreases have a minor effect.
Modern concepts further incorporate precautionary saving motives. They assume that people who face greater uncertainty about their living standard would save more. For those individuals the level of saving would exceed standard life-cycle saving, while people in safe working contracts would generally save less or start retirement saving later. In addition, general risk preferences would shape saving behavior: risk-friendly people would tend to spend more, whereas risk-averse individuals save a greater percentage of their income.
Moreover, bequest motives are integrated in the life-cycle saving concept. It is assumed that people do not only save for their own but also for future generations. Hence, saving can potentially reach higher levels than life-cycle models would suggest when assets are accumulated forthe purpose of intergenerational transfers, in particularwithin higher lifetime-income groups.
Finally, the recently discussed ‘wealth effect’ theory should be introduced as a position more in line with Keynes and Duesenberry than with Modigliani and Friedman. That is, household would adjust their saving and consumption decision to current changes in income. Temporary increases in income through capital gains on assets would reduce saving whereas current-period decreases in incomes would lead to reduced spending. Moreover, it is assumed that not only actually generated income influences the saving decision, but also the perceived income. That is, households would believe themselves to be richer in view of rising values of their assets in a temporarily environment of economic prosperity.
On balance, the wealth effect approach would lead to a different conclusion than the life-cycle theories: saving is determined by current (actual or perceived) increases in income and wealth ratherthan smoothed out constantly over a lifetime.
Before discussing the empirical finding, it is important to note that generally saving is regarded as a ‘noisy’ variable due to the variety of factors that potentially influence saving and the impact of saving on other macroeconomic spheres. Moreover, with respect to households and their diverging preferences and characteristics it may be more valuable to figure how people save than attempting to fully explain why people save.
Some households simply may be unwilling to save, preferring current spending over saving to secure the future. Not everyone is a life-cycle saver. Target-savers may save temporarily towards particular goals, but feel no need to save more after a fixed income level is reached. Low-income households may feel that they cannot afford to save while rich households may feel encouraged to consume more than to save, or vice versa. Hence, neither the life-cycle hypothesis, nor one of the alternative concepts has been fully verified with empirical data so far. While some evidence may speak for each motive, there exists no theory that ultimately explains personal saving behavior. However, regards saving patterns across population groups, the evidence merely supported consumption-smoothing and permanent-income lifecycle approaches over current- or relative- income hypothesis until very recently.
In the light of the theoretical implications examined so far, putting aside individual time and risk preferences, personal saving is assumed to be mainly determined by demographics, income and wealth. Following, summary results on these influential factors from empirical studies on SCF, CEX, and PSID data, to the very most part gathered in the 1980s, will be reviewed. In order to not get lost in methodological issues at this early stage, I simply note that the evidence presented was mainly derived from measuring saving either as income minus consumption (CES) or change in wealth (SCF and PSID) whereas unrealized capital gains were not attributed to saving in CES and SCF.
Speaking of demographics, most empirical studies on U.S. micro-level data suggest positive saving rates across all age groups. There is also evidence for a generally hump-backed shape of the age distribution of saving. In line with the lifecycle hypothesis, most observed household saving rates tent to increase with age, reached their peak towards retirement, and dropped thereafter.
A further demographic or rather socio-economic factor is family composition. Micro-level analysis basically revealed that family size and marital status affect saving rates. Findings indicate highest saving rates for married households without children, modest saving for couples with children, and by far lowest saving among lone parents. Generally, there is evidence that marriage and constant family structure fosters household saving. However, in the context of age and family structure, it is to remind that the household as the unit of analysis can be comprised of several individuals in different age groups and with diverging saving preferences.
As regards income and wealth, most 1980s micro-level surveys point to a positive correlation in the long term: a large proportion of total saving is due to those households sustaining in the upper part of the income and wealth distribution. In or- derto identify consumption smoothing patterns and impatient reactions to increases or decreases in current income, researchers controlled for those variables clearly related to permanent income. Implying the widely proven positive correlation of education and long-term income, it is revealed that saving is also higher with higher educational degrees. Further, the assumed negative correlation of certain ethnic backgrounds and income levels can be applied. There is reasonable evidence that Afro American households tend generate less savings along with lower income levels than white families.
The general picture is that saving increases with long-term income, whereas in particularthose households with stable living standards save more or rather put more effort in saving. After mainstream lifecycle theory and corresponding evidence, those households with high lifetime earnings (due to educational attainment and ethnic background) and with permanent family structures carefully accumulate savings over a lifetime in order to adjust for potential losses, to purchase further assets, to pay back loans, and to finance retirement or even future generations.
However, the recent massive overall decline of the U.S. household saving rate poses a challenge to this stylized pattern of personal saving. In the next section, I will show why the mainstream life-cycle concept has experienced reasonable criticism in the face of declining saving rates.
The aggregate U.S. household saving rate has declined significantly since the early 1980s. Figure 1 shows the NIPA and FFA annualized household saving as a ratio to disposable household income from 1950 to 2007.
The chart reveals a significant negative change in household saving. From I960 to 1980, household saving rates were comparably stable. As regards NIPA, rates were mainly fluctuating between 8 and 10 percent of disposable income, even experiencing a modest overall increase from roughly 7.5 percentage points in the early 1950s to over 10 percent in the early 1980s. The NIPA household saving rate averaged 8.3 percent over the 1960s and rose to an average of 9.6 percent in the 1970. Under FFA, saving rates were even higher, ranging between 9 and 14 percent of income for the period from 1950 to 1980.
Figure 1: NIPA and FFA U.S. householdsaving-to-income ratio, 1950-2005.
illustration not visible in this excerpt
Source: Goldsmith (1955) and U.S. Bureau of Economic Analysis (2010).
After decades of high and stable saving rates, from 1985 on, household saving dropped by a factor of 5 within twenty years. In the early 1980s, the NIPA household saving-income ratio peaked at upto 11 percent and the FFA rate even rose to over 14 percent of income, whereas by the end of the decade household saving had declined to 7.3 percent. After a short recovery towards 1991, reaching values of about 8 percent, saving started to decline steadily in the 1990s and in the new millennium. In the late 1990s, both NIPA and FFA measures averaged below 4.7 percent and reached even lower values at the turn of the millennium, when the saving-income ratio was roughly 2.2 percent, indicating a new all time low since the Great Depression. FFA even reported zero and slightly negative values not estimated since 1932/33.
To further illustrate the historical singularity of this event, Figure 2 traces the development of saving rates from 1900 to 2005 as a smoothed average of the NIPA measure and historical saving rates derived from the comprehensive work of R.W. Goldsmith.
illustration not visible in this excerpt
Speaking of the first part of the 19th century, the pattern of saving is comparatively easy to figure as external macroeconomic and political events shaped saving behavior. Saving rates rose to over 14 percent of income during the outbreak of World War I and to over 18 percent in the face of World War II before rebounding to normal levels about 8 percent thereafter. Regarding the interwar period, the Black Friday and the Great Depression dropped saving to a historical low about 2 percent of income as the overall economic situation in terms of rising unemployment, deflation, and stock market crash worsened. After the turbulences of the first part of the century, saving rates leveled off to constant averages about 8to10 percent of income and slowly increased in the generally rigid and stable overall economic regime from the 1950s to 1980. However, the huge savings drop from the mid 1980s on in the absence of severe political struggles and economic shocks or rather in times of expansion seems paradox also with regard to the historical development.
Recently, a great number of studies offer a variety of possible explanations for the sharp drop in saving by American households. To sum up, the eight most frequently discussed issues will be presented here.
a) Changes in the age structure
b) Cohort effects
c) Changes in the family structure
d) Increased government insurance
e) Less liquidity constrains trough rising credit availability
f) Privatization of pension wealth
g) Changes in income and wealth (and in the distribution of income and wealth)
h) Capital gains on assets such as housing and stocks
It is obvious that all potential explanations are somehow interwoven, which hardly allows for a separate analysis of each without controlling for the impact of the rest. Nevertheless, I will group the issues into three major categories: demographics (a-c), ‘external’ factors (d-f), and income and wealth (g-h), which will be reviewed in the following section.
 See Browning and Lusardi (1996) and Parker (1999) for a general overview of the suggested explanations.
 Bosworth and Anders (2008), 1.
 Juster et al. (2004), 1.
 Gale and Sabelhaus (1999), 4ff.
 Hence, the FFA saving rate is typically somewhat higher. See GAO (2001), 19. For a detailed discussion on the differences between NIPA and FFA see Maki and Palumbo (1991) and Verma and Lichtenstein (2001).
 GAO (2001), 18ff.
 Gale and Sabelhaus (1999), 25., Parker (1999), 317.
 For a detailed discussion on the comparability of FFA, PSID, and SCF wealth data, see Curtin, Juster and Morgan (1989).
 The standard reference is Keynes (1936).
 See Browning and Lusardi (1996), 1797.
 Duesenberry (1949), 3., Alvarez-Cuadrado,
 See Alvarez-Cuadrado and Van Long (200) for a discussion on the relative income hypothesis and its relation to the lifecycle hypothesis.
 The standard references are Modigliani and Brumberg (1954) and Ando and Modigliani (1963), see also Irvine and Wang (2001).
 See Modigliani and Brumberg (1954).
 The standard reference is Friedman (1957).
For possible origins of individual risk preferences see Cronqvist and Siegel (2010), Iff, Hintermaier and Koe- niger (2008), 22ff.
 For a discussion on the explanatory power of bequest see Berneim (1992). For the correlation of capital transmission between generations and wealth levels see Dynan, Skinner and Zeldes (2004) and De Nardi (2001).
 See Juster Smith and Stafford (1999), Juster et al. (2000), Juster et al. (2004) and Maki and Palumbo (2001).
 Avery, Kennickell (1991), 432.
 See Browning, Lusardi (1996).
 GOA (2008), 24f.
 Avery, Kennickel (1991), 422, Bosworth, Burtless, Sabelhaus (1991), 199f.
 Attanasio (1993), 17ff., Curtin, Justin and Morgan (1989), 523., Bosworth, Burtless, Sabelhaus (1991), 214, Deaton and Paxson (1999), 3, Parker (1999), 340. These authors find general support for Modigliani's theory but identify deviations with regard to different levels of education, race, income and wealth, and pension schemes. Moreover, it is assumed that the 'hump' towards retirement is somewhat flatter than suggested in theory (Attanasio 1993)). Lusardi (2001) points to the fact that many families still arrive at retirement with no money at all, also due to insufficient planning. Furthermore, Shorrocks (1975) and Jianakoplos, Menchik, and Irvine (1989) first demonstrated that cross-sectional analysis cannot be taken as a full confirmation of the lifecycle theory as cohort effects and mortality rates distort mean savings comparisons. See Wolff (1999) for more details.
 Browning and Lusardi (1996), 1821ff.
 ibid, 1815.
 Avery, Kennickell (1991), 422ff., Bosworth, Bell (2005), 16. Deaton (1989), 1, Lusardi (2001) 3.
 See, for example, Gale and Pence (2006), 207.
 Attanasio, DeLeire (1994), 13.
Gittleman and Wolff (2004), 223ff.
Gale and Sabelhaus (1999).
 Goldsmith (1955) measured trends in the saving-to-income ratios of the total U.S. population and different sectors over the first part of the 20th century. For details on methodology and scope see Goldsmith (1955), pp. 23ff.
 See Goldsmith (1955), 47.
Der GRIN Verlag hat sich seit 1998 auf die Veröffentlichung akademischer eBooks und Bücher spezialisiert. Der GRIN Verlag steht damit als erstes Unternehmen für User Generated Quality Content. Die Verlagsseiten GRIN.com, Hausarbeiten.de und Diplomarbeiten24 bieten für Hochschullehrer, Absolventen und Studenten die ideale Plattform, wissenschaftliche Texte wie Hausarbeiten, Referate, Bachelorarbeiten, Masterarbeiten, Diplomarbeiten, Dissertationen und wissenschaftliche Aufsätze einem breiten Publikum zu präsentieren.
Kostenfreie Veröffentlichung: Hausarbeit, Bachelorarbeit, Diplomarbeit, Dissertation, Masterarbeit, Interpretation oder Referat jetzt veröffentlichen!