Perceived Income Adequacy Among Older Adults in 12 Countries: Findings From the Survey of Health, Ageing, and Retirement in Europe

Abstract

Purpose: To validate a survey research measure of subjective income, as measured by perceived income adequacy, in an international context. Design and Methods: The study population comprised persons aged 50 years and older in 12 countries from the Survey of Health, Ageing and Retirement in Europe (n = 28,939). Perceived difficulty in making ends meet was regressed on sociodemographic variables, economic indicators, health status measures, and expectations regarding one's financial future. Country differences were also controlled. Results: The findings confirm a multidimensional explanation of perceived income adequacy but also point to the primacy of objective economic indicators in predicting household financial distress. Respondents aged 80 years and older report less financial difficulty. Poor health status and pessimistic financial expectations also predict greater household financial distress but to a lesser degree. Implications: Self-rated economic status is a robust indicator of financial capacity in older age and can be used by practitioners to gain meaningful information. However, practitioners should keep in mind that the oldest-old may underestimate financial difficulties.

Perceived income adequacy is an integral part of one's economic well-being at any age and an especially important indicator for understanding the financial capacity of older people. This is because incomes tend to decline in late life, due to retirement, and health-related expenses tend to rise, insurance coverage notwithstanding (Cook & Settersten, 1995; Stoller & Stoller, 2003). Yet, despite these trends, older people frequently express high levels of income satisfaction, sometimes unreasonably so (George, 1992). It is necessary, therefore, to confirm the validity of perceived income adequacy as a reliable measure of economic status in late life.

Study of subjective income is an important undertaking for social policy and gerontological practice. There is ongoing debate as to the level of income that is required to manage one's household in late life. Validation of a subjective economic status indicator can contribute to this discussion. It can also add to better assessment of the capacity of older clients to meet their financial needs and, as such, to estimate their ability to maintain independent community-based living.

The present analysis utilizes data from the first wave of the Survey of Health, Ageing and Retirement in Europe (SHARE), to validate a measure of subjective household financial distress (Börsch-Supan, Hank, & Jürges, 2005). The comprehensive, multidisciplinary, and international database that is provided through SHARE allows consideration of a wide range of objective and subjective factors in relation to perceived income adequacy among older adults. The database also allows cross-national, ex-ante harmonized comparisons among the 12 countries that participated in the first wave of the survey, from Sweden in the north to Greece and Israel in the south. As such, it allows examination of the correlates of subjective income while controlling for unique aspects of national context.

Perceived Income Adequacy

Perceived income adequacy has been addressed in the literature in various ways. One approach is based on the minimum income needed to live, that is, a kind of subjective poverty line (Stanovnik, 1992). Analysis of the consensual poverty line in Sweden found, for example, that a fifth of all households had incomes that were considered by most people as too low (Hallerod, 1995), although in some cases the incomes in question were higher than the eligibility cutoff for social assistance. In contrast, a Russian study revealed that perception of the minimum income needed to make ends meet tended to decline along with a decrease in real income (Milanovic & Jovanovic, 1999).

Another approach is to look at perceived income adequacy as the income one requires to live at the level that one desires. A number of studies suggest that this broader interpretation of income adequacy is influenced by several factors above and beyond the effect of objective economic measures (Hazelrigg & Hardy, 1997; Wilkie, Peat, Thomas, & Croft, 2007). These include relative perceptions of income adequacy (comparisons to others) and one's assessment regarding one's future needs. The variability inherent in subjective evaluation of income adequacy has led Ballantyne and Marshall (2001) to counsel against the use of this measure as a means to determine income redistribution policy. In contrast, Saunders and Matheson (1993) point out that the social meaning of income levels and their correlates have implications for public policy and should be addressed. More research is needed, therefore, to understand the nature and the implications of this self-rated economic status measure.

However, studies that have addressed perceived income adequacy tend to employ the construct as a predictor of other outcome measures. For example, subjective income has been examined in relation to self-rated health (Cairney, 2000), life satisfaction (Coke, 1992), depressive symptoms (St. John, Blandford, & Strain, 2006), health inequalities (Wildman, 2003), functional adaptation and home modification (Mathieson, Kronenfeld, & Keith, 2002), and even mortality (Blazer, Sachs-Ericsson, & Hybels, 2005). Few studies address perceptions of income adequacy as the outcome variable of interest. One such example is a study in which household income, health status, age, and migrant status were found to be the main predictors of perceived income adequacy among older Americans (Stoller & Stoller, 2003).

Among the cited correlates of perceived income adequacy, the literature notes age as a leading sociodemographic factor; the older one is, the less one reports difficulty in managing financially. Moreover, the association between age and income adequacy is maintained even when controlling for income and other variables (Hazelrigg & Hardy, 1997; Herzog & Rogers, 1981; Palmore, Fillenbaum, & George, 1984; Stanovnik, 1992). Two explanations have been suggested for this phenomenon. The first maintains that general expenses diminish in late life, allowing older people to manage with lower incomes. The second explanation points to the accumulation of wealth that offsets a decrease in income in old age, allowing elderly individuals to maintain their standard of living from alternative sources (George, 1992; Smeeding, 1990; Tacchino & Saltzman, 1999). Yet, another possible explanation might stem from cognitive dissonance theory; that is, having lower incomes may lead older people to change their interpretations as to how much is needed to get by.

Gender is less consistently associated with perceived income adequacy. Some studies report a significant association (Chan, Ofstedal, & Hermalin, 2002; Danigelis & McIntosh, 2001). In others, however, the gender effect disappears when controlling for other variables, particularly economic indicators (Hazelrigg & Hardy, 1997; Hsieh, 2003). Education may have a similar effect insofar as higher education facilitates greater earning capacity over the life course. Employment status may also contribute directly to income in older age, although adequate pension benefits could offset the decrease in income due to retirement.

Objective economic indicators have been found to predict perceptions of income adequacy above and beyond the effects of sociodemographic background characteristics. For example, a Dutch study found an association between the perceived minimal income needed to meet expenses and several objective economic indicators that included poverty level; ratio of income to food expenses, fixed costs, and overall expenses; and minimum salary (Hagenaars & de Vos, 1988). Similar findings have been reported elsewhere (Danigelis & McIntosh, 2001; Hsieh, 2003). In a study of perceived income adequacy, Hazelrigg and Hardy (1997) addressed both absolute and relative measures of financial status in relation to the subjective economic status outcome.

The literature also reports an association between functional health and subjective economic status, although the nature of this association requires greater clarification. Depression is strongly associated with perceived income inadequacy, but the association may work in both ways. That is, whereas poor mental health may lower one's assessment of being able to make ends meet, poor financial status might equally lead to decreased morale (Keith, 1993; Krause, 1997). As for physical disability, an inverse association with perceived income adequacy has been found, but it is often outweighed by the effects of other variables (Angel, Frisco, Angel, & Chiriboga, 2003).

Another relevant area involves expectations that individuals may have regarding their future earnings (Das & Donkers, 1999; Dominitz & Manski, 1997). A British study found in this regard that optimistic expectations resulted in greater risk taking in relation to loan acquisition (Brown, Garino, Taylor, & Price, 2005). The study showed, moreover, that this behavior continued over time even after evidence was received that optimistic financial expectations were not justified. In contrast, a Dutch analysis reveals that expectations regarding one's financial future are influenced by one's prior financial history (Das & van Soest, 1999). In that study, persons whose incomes had decreased tended toward pessimistic expectations regarding the future. Those whose incomes had increased tended toward a more optimistic appraisal of their upcoming financial dealings.

This review underscores the multifaceted nature of perceived income adequacy. Based upon the literature, we propose a multidimensional model through which to examine this subjective economic measure. The model has four principal components: (a) sociodemographic background characteristics that shape economic capacity in older age, including employment status; (b) objective measures of income and wealth, including absolute and relative indicators; (c) functional health status as a reflection of need and possible dependency; and (d) subjective expectations regarding one's financial future. The current analysis attempts to clarify the relative importance of the components of the analytic model, that is, the predictive strength of the explanatory factors. We look especially at the contribution of objective economic indicators in an effort to understand whether perceived income adequacy is a valid measure of financial capacity. We also examine other potential correlates, particularly age, to appreciate the conditions under which the perception of income adequacy may change in late life.

Methods

The SHARE data include persons age 50 years and older from a dozen countries, and their spouses of any age. Based on probability samples, SHARE represents the community-dwelling older population. In total, 31,115 persons were queried in the first wave by means of computer-assisted face-to-face interviews. The current analysis focuses upon respondents aged 50 years and older. Respondents with missing values on at least one measure were excluded. The resultant sample numbered 28,939 individuals (93% of the total number of observations). As in the Health and Retirement Study and the English Longitudinal Study of Ageing, each SHARE household designated one eligible household member as its financial informant. To allow data analysis on the individual level, we copied the data provided by the financial informant to the partner in the household.

Variables

The dependent variable was derived from the question, “Thinking of your household's total monthly income, would you say that your household is able to make ends meet?” The answer scale was composed of four categories ranging from with great difficulty to easily. The values were recoded into a dichotomous variable with the value of 1 for respondents who reported subjective economic hardship, that is, difficulty in making ends meet.

Due to the multifaceted nature of perceived income adequacy, we looked at explanatory variables from four main groups: sociodemographic background, objective measures of current economic status, health status, and expectations regarding one's financial future. The sociodemographic group included four variables: (a) age; (b) gender; (c) level of education, harmonized for international comparison by recoding raw values into United Nations Educational, Scientific, and Cultural Organization's International Standard Classification of Education (1997); and (d) self-reported employment status. The latter differentiated between the employed/self-employed, retired, unemployed, sick/disabled, and homemakers.

Economic status was measured by standardized monthly household income, individual net worth, and relative income. The income variable was based on self-reports of gross individual income from employment, pensions, transfers, and other sources (Brugiavini, Croda, Paccagnella, Rainato, & Weber, 2005). Respondents who did not remember exact amounts were aided in their estimations by a series of unfolding bracket sums. Those who gave no amount at all were assigned values through hot-deck imputation, a commonly used item nonresponse adjustment procedure that employs current survey responses to substitute for missing data. The resultant sums were adjusted for relative purchasing power parity (PPP) within the participating SHARE countries and standardized by the household size square root to get the equivalent disposable income, per standard person. The net worth variable was based on the combined value of primary residence, bank accounts, stock holdings, and other sources, after deducting financial liabilities (Christelis, Jappelli, & Padula, 2005). Purchasing power parity adjustments were performed here as well. Both of these variables were recoded into three equal-size groups. The final economic measure—relative income—was a dummy variable that reflected one's individual income in relation to the overall sample median. Scores of 1 reflected incomes at or above the median.

The health status measures consisted of two dichotomous variables. Physical activity limitations were measured on a scale similar to the Physical Activity Scale employed in the Yale Health and Aging Project (Cornoni-Huntley et al., 1985). Respondents indicated whether they had difficulty in pulling a large object, stooping, lifting, reaching, and picking up a small coin. In the present analysis, a value of 1 reflected any mentioned physical limitations. Mental health was measured on the EURO-D scale, which was developed to facilitate cross-cultural research into late-life depression in Europe. It is based on self-report of 12 symptoms, such as excessive guilt feelings, trouble sleeping, and irritability (Castro-Costa et al., 2007). A value of 1 on the dichotomous measure employed here reflected a EURO-D scale score above the clinical cutoff point of three symptoms.

The fourth group of variables considered respondents’ expectations concerning their financial futures. The variables are based upon the percentage estimation that one's standard of living will improve or worsen, 5 years from the time of interview. Based on this estimation, we calculated two dummy variables with the value of 1 for estimations of chances greater than 50% for improvement (optimism) or for worsening (pessimism). Lastly, to account for national differences in perceived income adequacy, we controlled for country of residence.

Analysis

The analysis proceeded in three stages. First, we looked at the univariate description of the study variables. Next, we examined the bivariate associations between each of the independent variables and perceived income inadequacy using the chi-squared goodness-of-fit test. In the third and final stage, we used a multivariate logistic regression to test the association between each predictor and the outcome variable, considering all other variables. We also executed the multivariate regressions separately by age group.

Two statistics are discussed in the multivariate analysis. The B coefficient shows the change in the predicted logged odds of financial distress for a 1-unit change in the explanatory variables. B's asterisks flag statistically significant associations between the predictor variables and the outcome measure. The odds ratio (OR), the exponentiation of the B coefficient, informs us as to the relative likelihood that people with a certain characteristic will experience financial distress. Thus, for example in Table 2, the OR for respondents with a low level of wealth is 3.11. This implies that these respondents were more than three times as likely to perceive their income as inadequate as respondents with a high level of wealth (the reference category). Conversely, an OR of less than 1.0 reflects a lower likelihood of perceived financial inadequacy. Table 2 indicates, for example, that respondents aged 80 or older had an OR of 0.47. This actually means that they were more than twice less likely to report subjective economic hardship (1/0.47 = 2.13). The findings regarding ORs lower than 1.0 are reported in this manner in the following text.

Table 2.

Likelihood of Reporting Difficulty in Making Ends Meet (logistic regression)

Variable	Category	B	Odds ratio	95% Confidence interval
Gendera	Women	−0.08*	0.92	0.86–0.98
Ageb (years)	60–69	−0.16***	0.85	0.78–0.92
	70–79	−0.34***	0.71	0.65–0.79
	80+	−0.76***	0.47	0.41–0.53
Educationc	Low	0.50***	1.65	1.51–1.79
	Medium	0.18***	1.20	1.10–1.31
Employment	Retired	0.17***	1.18	1.08–1.30
Statusd	Unemployed	0.84***	2.32	1.98–2.73
	Sick/disabled	0.73***	2.08	1.77–2.45
	Homemaker	0.16**	1.18	1.06–1.31
Net worthe	Low wealth	1.14***	3.11	2.89–3.36
	Medium wealth	0.43***	1.54	1.44–1.66
Incomef	Low income	0.66***	1.94	1.72–2.18
	Medium income	0.18***	1.20	1.10–1.31
Relative incomeg	≤Median	0.37***	1.45	1.32–1.59
EURO-Dh	>3	0.48***	1.61	1.51–1.72
Disabilityi	>0	0.29***	1.33	1.25–1.42
Pessimismj	>50%	0.45***	1.57	1.47–1.68
Countryk	Austria	0.24*	1.27	1.03–1.58
	Belgium	0.54***	1.71	1.40–2.08
	Denmark	0.05	1.05	0.83–1.31
	France	0.71***	2.03	1.66–2.48
	Germany	0.23*	1.26	1.03–1.55
	Greece	2.23***	9.32	7.57–11.47
	Israel	2.09***	8.08	6.58–9.94
	Italy	2.02***	7.56	6.14–9.28
	Netherlands	−0.16	0.85	0.69–1.05
	Spain	1.54***	4.66	3.78–5.74
	Sweden	0.01	1.01	0.82–1.25

Notes: ^aReference: men.

^bReference: 50–59 years.

^cReference: high education.

^dReference: employed.

^eReference: high wealth.

^fReference: high income.

^gReference: >median.

ⁱReference: no physical disability.

^jReference: ≤50%.

^kReference: Switzerland.

*p < .05. **p < .01. ***p < .001; Nagelkerke r² = .356.

Results

On the whole, 38% of respondents felt that their households managed with some degree of difficulty. Table 1 presents the frequency distributions of the independent variables. There were slightly more women in the sample. Respondents aged 50–69 years made up more than two thirds of the sample, but only some 29% of the sample were employed. One half of all respondents had low education levels. In terms of health, a bit more than a third of the respondents reported having one or more physical activity limitations and one quarter of them were depressed. About one quarter of the respondents were pessimistic regarding their financial future, and only 13% were optimistic. The income and wealth variables divided proportionately in country tertiles or according to the European median, by design.

Table 1.

Description of the Study Variables and Their Association With Perceived Difficulty in Making Ends Meet (frequency distributions and cross-tabulations)

	Univariate statistics			Difficulty making ends meet
Variable	Category	n	%	%	χ2
Gender	Men	13,223	45.7	36.2	52.78*
	Women	15,715	54.3	40.3
Age (years)	50–59	10,706	37.0	38.1	7.60†
	60–69	9,281	32.1	37.8
	70–79	6,281	21.7	39.6
	80+	2,670	9.2	37.2
Education	Low	14,411	49.8	47.9	1,291.28*
	Average	8,600	29.7	33.4
	High	5,927	20.5	23.2
Employment status	Employed	8,305	28.7	30.5	835.60*
	Retired	14,245	49.2	37.1
	Unemployed	948	3.3	59.1
	Sick/disabled	976	3.4	57.0
	Homemaker	4,464	15.4	49.3
Net worth	Low	9,481	32.8	53.2	1,919.41*
	Average	9,616	33.2	38.9
	High	9,841	34.0	23.6
Income	Low	9,454	32.7	60.0	3,432.88*
	Average	10,276	35.5	34.8
	High	9,208	31.8	20.4
Relative income	<Median	14,008	48.4	49.0	1,352.39*
	≥Median	14,930	51.6	28.6
EURO-D	No	21,538	74.4	33.4	924.77*
	Yes	7,400	25.6	52.9
Disability	No limitations	10,804	37.3	45.3	519.61*
	1+ limitations	18,134	62.7	54.7
Pessimism	No	21,521	74.4	37.0	91.87*
	Yes	7,417	25.6	43.1
Optimism	No	25,281	87.4	38.4	1.39
	Yes	3,657	12.6	39.3
Country	Austria	1,824	6.3	26.5	4,773.27*
	Belgium	3,625	12.5	29.3
	Denmark	1,579	5.5	19.3
	France	2,807	9.7	34.9
	Germany	2,877	9.9	26.1
	Greece	2,579	8.9	70.4
	Israel	2,367	8.2	61.5
	Italy	2,482	8.6	65.8
	Netherlands	2,723	9.4	20.2
	Spain	2,259	7.8	58.6
	Sweden	2,893	10.0	20.4
	Switzerland	923	3.2	18.3

Note: †p = .055; *p < .001.

The bivariate associations between perceived income adequacy and the independent variables also appear in Table 1. All the study variables, except for optimistic financial expectations, were related to the subjective economic outcome measure (age was marginally related). Women reported greater difficulty making ends meet, as did respondents with lower education. Those who were employed had less difficulty, as did those with higher incomes and net worth. Respondents who were pessimistic about their financial futures sensed greater current financial difficulty as well. Finally, those with mental or physical limitations reported greater household financial difficulty.

A separate analysis (not shown here) also looked at respondents’ consumption levels in relation to perceived income adequacy. The sum total of household expenses for food, utilities, goods, and services was divided into tertiles and cross-tabulated with perceived difficulty in making ends meet. The analysis revealed no association between the two. Consequently, this variable was not considered in the present inquiry. Table 1 also includes the rates of difficulty making ends meet by country. The range was considerable—from a low of 18% in Switzerland to a high of 70% in Greece. Given these large differences, the multivariate analysis controlled for the country variable. Switzerland served as the reference category.

The multivariate results appear in Table 2. The model accounted for some 36% of the variance. First of note is the age effect that appeared when all the study variables were considered. Respondents aged 60–69 years were slightly less likely to report difficulty in making ends meet than those aged 50–59 years, respondents aged 70–79 years were even less likely, and those aged 80 years and older were much less likely to report financial difficulty. Women were somewhat less likely to report financial difficulty when the other variables were controlled. Lower education and all the categories of nonemployment were associated with perceived financial difficulty, particularly among those who were unemployed or disabled.

The economic indicators were all strong predictors of perceived income adequacy. Respondents with low incomes were almost twice as likely to report having financial difficulty as those with high incomes, all else considered. Respondents having low net worth were more than three times as likely as respondents with a high degree of net worth. Relative income was also a significant predictor—those with incomes lower than the sample median were a bit more likely to report difficulty in making ends meet. In addition, financially pessimistic respondents were somewhat more likely to perceive their current household incomes as inadequate. As for health, both measures retained their previous bivariate association but to a more moderate degree. Depression emerged as a stronger predictor of poor subjective economic status than the physical limitation measure. Finally, significant country differences remained even after controlling for the effects of the other variables in the analysis.

To more closely understand the relationship of age and perceived income adequacy, we ran the same regression separately for each age group, controlling once again for country. Figure 1 presents salient selected findings from these analyses. The results for the four separate age categories are grouped by the predictor variables for purpose of comparison; thus, there are four age bars for each variable. Almost all the associations were significant (a transparent bar indicates the few that were not). The two strongest predictors of financial distress were the objective financial indicators: wealth and income. However, the figure also shows that the strength of association among these variables drops considerably among the oldest-old, with the income variable losing its significance. Pessimistic financial expectations were also significant predictors of household financial difficulty for all age groups but more so among respondents aged 70 years and older.

Figure 1.

Predictors of financial distress by age groups^a: unstandardized logged odds coefficients (B).

Note: ^aAdditional variables in the analyses (but not shown in the figure) included country, medium education, employment status, medium wealth, and medium income. Nagelkerke pseudo-R²: age 50–59 years, .36; age 60–69 years, .38; age 70–79 years, .38; age 80+ years, .29; p < .01.

A second grouping of predictor variables included low education, depression, and physical disability. The effect of these variables on perceived household financial difficulty did not vary much by age group. Low relative income was a more modest predictor of household financial distress with no major age differences discerned. Finally, gender was significant only among the youngest age group; women aged 50–59 years reported less financial distress than men of the same age.

Discussion

This analysis sought to better understand the nature of perceived income adequacy among older adults and to confirm its validity as a measure of economic status in late life. The study population comprised persons aged 50 years and older in 12 countries from the first wave of SHARE. The results revealed that on the whole, the primary predictor of income adequacy was an objective economic measure—net wealth. The second strongest set of predictors comprised two employment status indicators (the unemployed and the disabled) and another objective economic measure—lowest income grouping. The main findings from the analysis thus underscore the economic basis of perceived income adequacy. Stated differently, a sense of difficulty in making ends meet among older adults reflects lower financial means and lesser earning capacity.

The results of the inquiry also showed that respondents aged 80 years and older reported significantly less financial difficulty when compared with the 50-year-olds, an effect similar in strength to that of the second most influential group of predictors previously mentioned. (Persons aged 60–79 years also reported less difficulty but to a more modest degree.) Thus, the strongest age effect on perceived income adequacy was observed among the oldest-old. This association exists above and beyond the respective effects of a wide range of objective economic indicators, health measures, and future financial expectations.

Poor health status and pessimistic financial expectations also predicted greater household financial distress but to a lesser degree than the objective economic indicators. Low education and gender were also related. Above and beyond the effects of the variables in the explanatory model, the current analysis also uncovered significant country differences in perceived income adequacy among the older population. Analyses of these country differences, their antecedents and effects, warrant additional inquiry.

To further scrutinize the association of age and subjective income, the conceptual model employed in the analysis was considered separately by age group. The results showed two main trends. First, the objective economic measures of wealth and income were the primary predictors of perceived income adequacy in all groups through age 79. Thus, the present study confirms that for the majority of older people, perceived income adequacy indeed reflects one's economic status. Subjective income is a measure that can and should be employed in research and in gerontological assessment.

However, a second trend also emerged in which the objective economic basis of subjective income weakened to some degree among the oldest-old. The results showed that wealth remained a major predictor of perceived income adequacy among those aged 80 years and older but income did not. Moreover, pessimistic financial expectations constituted the primary correlate of household financial distress among this group, all else considered. What do these findings imply regarding the use of subjective income indicators in research and in practice?

Gerontological literature has considered the linkages between objective and subjective appraisals in other areas, such as health and functioning (Bowling & Stafford, 2007; Saevareid, Thygesen, Nygaard, & Lindstrom, 2007). However, the findings from such studies are mixed. For example, a recent meta-analysis found that correlations of subjective health with physical and functional health were indeed lower in the old-old than in the young-old but that associations of subjective health with mental health were stronger (Pinquart, 2001). The varied sets of findings hint that the association between objective and subjective measures among the oldest-old might be mediated by age-related and other factors. Further research on this relationship is required, particularly on the association between objective and subjective economic indicators.

The practical implications that stem from the current analysis are twofold. First, we believe that indicators of perceived income adequacy are sufficiently robust for use in gerontological research and assessment. A global question about how older people manage financially can provide useful information in both survey questionnaires and intake inventories. Second, it seems that additional inquiry is required to better understand the economic status of the oldest-old who tend to underestimate financial difficulty. The results from the current analysis point to the need for applying anchoring questions in the study of perceived income adequacy among this age group.

Anchoring vignettes have recently been developed in a variety of disciplines to improve the validity and cross-cultural comparability of data from survey research (King & Wand, 2007). Such vignettes present hypothetical situations about which people are asked to react. Their responses to these situations provide an anchor or standard score by which their subjective evaluations can be weighted. This methodology has been found to improve the quality of comparison across different cultural groups (Kristensen & Johansson, 2008; Salomon, Tandon, & Murray, 2004). It may be efficacious to apply the methodology to comparison of perceived income adequacy across different age groups as well.

A limitation of the present study is its cross-sectional nature. Longitudinal data are needed to trace the effects of changes in the predictor variables on corresponding changes in perceived income adequacy. Such inquiry will be possible in the years to come, as additional waves of SHARE data are collected. A second limitation of the study reported here involves the question of consumption. Do older people feel they manage sufficiently because they spend less? Does having fewer resources lead them to reassess how much one needs to manage the household without difficulty? As mentioned earlier, preliminary analysis revealed no apparent association between total household expenses and perceived income adequacy in this sample. Nevertheless, the role of expenditures in relation to self-rated economic status requires greater elaboration in future analysis.

In conclusion, the present study validates older respondents’ subjective assessment of their financial situation as an important tool in understanding their actual financial status and points to the primacy of objective economic indicators in predicting household financial distress. As such, self-rated economic status can be said to be a robust indicator of one's financial capacity in older age. The study results also suggest that the oldest-old may overestimate their financial capabilities, above and beyond an impartial assessment of their means. Practitioners should keep this caveat in mind when very elderly clients tell them that they are able to make ends meet.

Funding

The SHARE data collection was primarily funded by the European Commission through the Fifth Framework Program (project QLK6-CT-2001-00360 in the thematic program Quality of Life). Additional funding came from the U.S. National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, Y1-AG-4553-01, and OGHA 04-064). SHARE data collection in Israel was funded by the U.S. National Institute on Aging (R21 AG025169), by the German-Israeli Foundation for Scientific Research and Development, and by the National Insurance Institute of Israel. The Israel Gerontological Data Center was established by a grant from the Ministry of Science, and is supported by the Ministry of Senior Citizens.