Subjective Well-Being, Income, Economic Development and Growth I have 20 minutes to speak In 1974, Richard Easterlin wrote a paper titled “Does Economic Growth Improve the Human Lot? Some Empirical Evidence”. This paper launched a literature of economists using subjective well-being data to assess the policy The first part of the title “Does Economic Growth improve the Human Lot” points to his ambitions The second part, “some empirical evidence” hints at his approach: He analyzes data from surveys asking people how happy they are, and this paper launched the economic analysis of happiness data. In recent decades, that literature has greatly expanded to hundreds of papers The “Easterlin Paradox” is easily the most cited finding in the economics of happiness, and the 1974 paper also launched a thirty-year research program for Richard Easterlin exploring the link between income and happiness. FLIP TO NEXT SLIDE HERE So, what is the Easterlin Paradox? 3 comparisons of the link between income happiness appear to give different answers Between rich and poor people in a society at a point in time: Strong evidence that income matters As far as I am aware, in every representative national survey ever done a significant bivariate relationship between happiness and income has been found.” - Easterlin 1974 Between rich and poor countries: No evidence, or weak evidence that income matters the happiness differences between rich and poor countries that one might expect on the basis of the within country differences by economic status are not borne out by the international data.” – Easterlin 1974 Within countries through time: Strong claim that economic growth does not change happiness “income growth in a society does not increase happiness” – Easterlin 1995 Title of Easterlin 1995: Will Raisng the Incomes of All Increase the Happiness of All? The juxtaposition of these three findings is the Easterlin Paradox Our task in this paper – joint with Justin Wolfers– is to revisit and revise the stylized facts about the links between income and happiness Our finding: Much of what we thought we knew, we don’t. Dan Sacks, Betsey Stevenson and Justin Wolfers Wharton School, University of Pennsylvania and NBER World Bank Workshop, NYU—October 8, 2010.

Implications of the Easterlin Paradox “Why do national comparisons among countries and over time show an association between income and happiness which is so much weaker than, if not inconsistent with, that shown by within-country comparisons?” –Easterlin (1974) Reference-dependent preferences Relative income matters [Other people’s consumption matters] Habit formation = hedonic treadmill [Other period’s consumption] Policy implications: Growth: “My results, along with mounting evidence from other time series studies of subjective well-being, do on balance undermine the view that a focus on economic growth is in the best interests of society.” –Easterlin (2005) Public finance: If preferences are interdependent  Pigouvian rationale for taxing labor supply / conspicuous consumption Within-country estimates: “in every single survey, those in the highest status group were happier, on the average, than those in the lowest status group” – Easterlin (1974) Between-country estimates: “the happiness differences between rich and poor countries that one might expect on the basis of the within country differences by economic status are not borne out by the international data.” – Easterlin (1974) National time series : “income growth in a society does not increase happiness” – Easterlin (1995) Sacks, Stevenson & Wolfers, Income and Happiness

Subjective Well-being “Subjective well-being refers to all of the various types of evaluations, both positive and negative, that people make of their lives. It includes reflective cognitive evaluations, such as life satisfaction and work satisfaction, interest and engagement, and affective reactions to life events, such as joy and sadness.” (Diener, 2005) Typical questions: Happiness “Taking all things together, would you say you are: very happy; quite happy; not very happy; not at all happy.” Life satisfaction “All things considered, how satisfied are you with your life as a whole these days?” [1=Dissatisfied – 10=Satisfied] Satisfaction ladder: “Here is a ladder representing the ‘ladder of life’. Let's suppose the top of the ladder represents the best possible life for you, and the bottom, the worse possible life for you. On which step of the ladder do you feel you personally stand at the present time? [0-10 steps].” If SWB is the “new” measure of welfare, and income is the “old” measure of welfare, I’m going to spend most of the rest of this talk comparing the new with the old. In particular: What is the relationship between SWB and income? Sacks, Stevenson & Wolfers, Income and Happiness

Research Question: What is the Relationship Between Subjective Well-being and Income? Types of comparisons 1970s view “Easterlin Paradox” 1990s view “Satiation” New view: Stevenson-Wolfers findings Within country: Rich v. poor people in a country Big effects Big effects for income <$15k - But not for the rich Strong effects: βwithin≈0.35 Between country: Rich v. poor countries Statistically insignificant effects interpreted to be zero GDP<$15k: Strong effects GDP>$15k: No effects Strong effects βbetween=0.35 National time series: Country when rich v. poor No effects (Japan, USA, Europe) No effects (Available data are for rich countries) Strong effects βtime series=0.35 International panel: Countries with fast v. slow growth Largely unexamined Strong effects βpanel=0.35 Implications Relative income determines well-being Relative income determines well-being… once “basic needs” are met” There’s no paradox (and never was) Policy conclusions De-emphasize growth Rich countries should de-emphasize growth and tax labor supply “My results, along with mounting evidence from other time series studies of subjective well-being, do on balance undermine the view that a focus on economic growth is in the best interests of society.” – Richard Easterlin (2005) Types of comparisons Within country: Between country: National time series: “income growth in a society does not increase happiness”. - Easterlin 1995 International panel data: Not really examined (long panels are relatively new) This paradox highlights the need for a focus on magnitudes rather than statistical significance. Easterlin found a statistically significant relationship between income and happiness within countries and a statistically insignificant relationship between countries. Yet the absence of evidence is not evidence of absence and (The difference between a statistically significant effect and a statistically-insignificant effect is not in itself statistically significant.) Yet I will show that new data helps with significance and a revisit of the old data reveals an inability to reject a positive relationship that is the same with and between countries. Our goal is simple to test these stylized facts and not to assess causality or revisit what SWB means. Before we go further however we need to know how subjective well-being is measured. (Don’t say) As in existing literature, we won’t ask whether it is GDP, broader measures of economic development such as institutions (political) or productivity that influence SWB. Sacks, Stevenson & Wolfers, Income and Happiness

Research Question: What is the Relationship Between Subjective Well-being and Income? Types of comparisons 1970s view “Easterlin Paradox” 1990s view “Satiation” New view: Stevenson-Wolfers findings Within country: Rich v. poor people in a country Big effects Big effects for income <$15k - But not for the rich Strong effects: βwithin≈0.35 Between country: Rich v. poor countries Statistically insignificant effects interpreted to be zero GDP<$15k: Strong effects GDP>$15k: No effects Strong effects βbetween=0.35 National time series: Country when rich v. poor No effects (Japan, USA, Europe) No effects (Available data are for rich countries) Strong effects βtime series=0.35 International panel: Countries with fast v. slow growth Largely unexamined Strong effects βpanel=0.35 Implications Relative income determines well-being Relative income determines well-being… once “basic needs” are met” There’s no paradox (and never was) Policy conclusions De-emphasize growth Rich countries should de-emphasize growth and tax labor supply “once a country has over $15,000 per head, its level of happiness appears to be independent of its income” – Richard Layard (2003) Types of comparisons Within country: Between country: National time series: “income growth in a society does not increase happiness”. - Easterlin 1995 International panel data: Not really examined (long panels are relatively new) This paradox highlights the need for a focus on magnitudes rather than statistical significance. Easterlin found a statistically significant relationship between income and happiness within countries and a statistically insignificant relationship between countries. Yet the absence of evidence is not evidence of absence and (The difference between a statistically significant effect and a statistically-insignificant effect is not in itself statistically significant.) Yet I will show that new data helps with significance and a revisit of the old data reveals an inability to reject a positive relationship that is the same with and between countries. Our goal is simple to test these stylized facts and not to assess causality or revisit what SWB means. Before we go further however we need to know how subjective well-being is measured. (Don’t say) As in existing literature, we won’t ask whether it is GDP, broader measures of economic development such as institutions (political) or productivity that influence SWB. Sacks, Stevenson & Wolfers, Income and Happiness

Research Question: What is the Relationship Between Subjective Well-being and Income? Types of comparisons 1970s view “Easterlin Paradox” 1990s view “Satiation” New view: Stevenson-Wolfers Within country: Rich v. poor people in a country Big effects Big effects for income <$15k - But not for the rich Strong effects: βwithin≈0.35 Between country: Rich v. poor countries Statistically insignificant effects interpreted to be zero GDP<$15k: Strong effects GDP>$15k: No effects Strong effects βbetween≈0.35 National time series: Country when rich v. poor No effects (Japan, USA, Europe) No effects (Available data are for rich countries) Strong effects βtime series≈0.35 International panel: Countries with fast v. slow growth Largely unexamined Strong effects βpanel≈0.35 Implications Relative income determines well-being Relative income determines well-being… once “basic needs” are met” There’s no paradox (and never was) Policy conclusions De-emphasize growth Rich countries should de-emphasize growth and tax labor supply Our goal is to revisit these stylized facts. Why? Theoretical implications: Reference-dependent preferences Yielding important policy implications (and big policy claims) The reconciliation of our findings with the previous literature rests on three observations New data: Longer time series (1946-2008); More countries (n=140) Statistical inference: Absence of evidence v. evidence of absence What are “big” versus small effects?  Focus on the magnitudes Before I go on: There are two things we won’t do: Assess causality Ask what subjective data “mean”. (I’ll leave that for Peter Hammond, my discussant, who has some wise insight.) Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness Why revisit the stylized facts? Theoretical implications: Reference-dependent preferences Yielding important policy implications (and big policy claims) What explains our new findings? New data: Longer time series (1946-2010); More countries (n=140) Statistical inference: Absence of evidence v. evidence of absence What are “big” versus small effects?  Focus on the magnitudes What we won’t do Assess causality: Happiness = β log(Income) Revisit what subjective well-being data “mean” Sacks, Stevenson & Wolfers, Income and Happiness

Outline: Assessing the Happiness-Income link Within-country cross-sectional comparisons USA All countries Between countries: In past and current data Multiple datasets For both happiness and life satisfaction No evidence of satiation National Time Series and International Panels Japan Europe World Values Survey Newly-available measures of subjective well-being We’ll start by looking at the relationship between subjective well-being and income seen within cross-sectional samples of countries because this is the relevant benchmark against which we want to compare the magnitudes of the relationships between well-being and income seen between countries and within countries over time Sacks, Stevenson & Wolfers, Income and Happiness

Within-Country Comparisons: USA “Taken all together, how would you say things are these days?” Family income Very happy Pretty happy Not too happy <$12,500 (bottom 10%) 21% 53% 26% $12,500-$49,999 25% 61% 13% $50,000-$149,999 40% 54% 6% $150,000 (top 10%) 45% 2% Source: U.S. General Social Survey, 2006 “When we plot average happiness versus income for clusters of people in a given country at a given time, we see that rich people are in fact much happier than poor people. It’s actually an astonishingly large difference. There’s no one single change you can imagine that would make your life improve on the happiness scale as much as to move from the bottom 5 percent on the income scale to the top 5 percent.” - Robert Frank (2005) Notice that there are no miserable millionaires As far as I am aware, in every representative national survey ever done a significant bivariate relationship between happiness and income has been found.” - Easterlin 1974 Start with data from the long-running US survey the General Social Survey which has been conducted since 1972 In the most recent data, 2006, we can see clearly that the rich are happier than the poor. (Describe what’s done before flipping slide for audience) We can look at this relationship for all of the years that the GSS has been run. To do so We run an ordered probit Happinessiyt=m*Income category*year We then plot these fixed effect estimates and examine the gradient of the slope of a regression of income on the fixed effects which is average happiness in each category normalized, the relationship is highly significant with a slop of .21 Sacks, Stevenson & Wolfers, Income and Happiness

Well-Being and Log(Income) βwithin ≈ 0.35 Discuss normalization Sacks, Stevenson & Wolfers, Income and Happiness Source: Daniel Sacks, Betsey Stevenson and Justin Wolfers (2010), “Subjective Well-Being, Income, Economic Development and Growth”

Within-country : Rich are happier than poor Without controls With controls Permanent Income Adjusted IV Sample size Gallup World Poll: Ladder question 0.236*** (0.014) 0.232*** 0.422 0.449 *** (0.027) 171,900 (126 Countries) World Values Survey: Life satisfaction 0.216*** (0.017) 0.227*** (0.037) 0.413 0.26*** (0.035) 116,527 (61 Countries) Pew Global Attitudes Survey: Ladder question 0.281*** 0.283*** 0.515 0.393*** (0.033) 32,463 (43 Countries) Notes: The table reports the coefficient on the log of household income, obtained from regressing standardized life satisfaction against the log of household income and country fixed effects using the indicated data set. Additional controls include a quartic in age, interacted with sex, plus indicators for age and sex missing. Our permanent income adjustment is to scale up our estimates by 1/0.55; see text for explanation. We instrument for income using full set of country×education fixed effects. We report robust standard errors, clustered at the country level, in parentheses. For further details on the standardization of satisfaction and the exact wording of satisfaction question, see the text. ***, ** and * denote statistically significant at 1%, 5% and 10%, respectively. Sacks, Stevenson & Wolfers, Income and Happiness

Outline: Assessing the Happiness-Income link Within-country comparisons USA All countries Between countries: In past and current data Multiple datasets Both happiness and life satisfaction No evidence of satiation National Time Series and International Panels Japan Europe World Values Survey βwithin ≈ 0.35 “the happiness differences between rich and poor countries that one might expect on the basis of the within country differences by economic status are not borne out by the international data.” – Easterlin, (1974) Now we want to turn to examining the magnitude of the relationship seen between countries. The purpose of the previous section was to establish a gradient seen within countries to which we could compare to that seen between countries Recall that previous research has claimed that the within gradient is much larger than the between gradient I’m going to start at the beginning and show you lots of data sets quite rapidly. Sacks, Stevenson & Wolfers, Income and Happiness

Early Cross-National Studies of Well-Being and GDP βbetween ≈ 0.35 Here are the earliest cross-national surveys of happiness Three observations: These are predominantly comparisons of developed countries with other developed countries, not of rich and poor countries The gray band highlights Easterlin’s interpretation “[w]hat is perhaps most striking is that the personal happiness ratings for 10 of the 14 countries lie virtually within a half a point of the midpoint rating of 5 [on the raw 0-10 scale]… A focus on statistical significance might lead one to reject an income-happiness link and Easterlin concluded: “The results are ambiguous… If there is a positive association between income and happiness, it is certainly not a strong one.” But the absence of evidence is not the same as evidence of absence and we can see that be looking at the point estimates: 1.44, .6, 1.3, .36, .21 Half are statistically insignificant and all would allow the possibility for a positive relationship between happiness and income that is similar to what is seen in the cross-section Most important—you cannot reject a null hypothesis that these coefficients are the same as those seen within countries. In no case can we reject a coefficient of 0.35 A precision-weighted average of these five regressions coefficients yields a coefficient of 0.45 Implication: There never was an Easterlin Paradox Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness βbetween ≈ 0.35 World Values Survey Mid-1980s: Still scope to debate whether there is a significant relationship, early wave consisted mostly of wealthy countries With the subsequent accumulation of data, that would be a very difficult position to maintain Pooling all four waves yields a happiness-satisfaction gradient of 0..40 (se=0.04) (.32(.05)) Differences in slope across panels are not statistically significant In the 1980s WVS almost no poor countries, added poor countries many of which were not representative but were rather subsets of wealthy, more educated members of those societies, these non representative countries are shown with the hollow squares (appendix in paper details which countries weren’t representative and how) Not including them yields coefficients of .5 .56 .46 .35 including them, .5, .2, .3, .3, all significant Sacks, Stevenson & Wolfers, Income and Happiness

Pew Global Attitudes Survey, 2002 βbetween ≈ 0.35 Pew data from 2002 relationship becomes clearer, correlation is .55, gradient is .22 The relationship becomes clearer in the Pew Data and note that there is no evidence of satiation—the lowess fit line indicates an increase in the slope for higher income countries And becomes even clearer in the 2006 Gallup World Poll=> flip to next slide Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness Gallup World Poll βbetween ≈ 0.35 Findings consistent with recent research by Angus Deaton Correlation of average happiness with average income =0.82 Then 2006 Gallup World Poll, correlation .82, gradient .4, highly statistically significant This data was designed by Gallup with the help of Danny Kahneman, Alan Krueger, Angus Deaton, and other experts in order to get the most reliable cross-country estimates of Subjective Well-being Again, looking at the locally weighted non-parametric line there is no evidence of satiation, if anything the relationship is stronger among the wealthier countries So let’s use the Gallup data to compare the well-being income gradient within countries to that seen across countries, so we clearly see the between countries gradient here We will take each country dot in this figure and add to it a line to show the within country slope Sacks, Stevenson & Wolfers, Income and Happiness

Between Countries: Rich Countries Are Happier Estimates of βbetween Microdata National Data Without controls With controls Sample size   Gallup World Poll: Ladder question 0.357*** (0.019) 0.378*** 0.342*** 291,383 (131 countries) World Values Survey: Life satisfaction 0.360*** (0.034) 0.364*** 0.370*** (0.036) 234,093 (79 countries) Pew Global Attitudes Survey: Ladder question 0.214*** (0.039) 0.231*** (0.038) 0.204*** (0.037) 37,974 (44 countries) βbetween ≈ 0.35 Notes: This table reports the coefficient on the log of per capita GDP, obtained from regressing standardized life satisfaction against the log of GDP, using individual data with and without controls, and using national-level data without controls, in the indicated data set. In the national-level regressions, we take the within-country average of standardized life satisfaction as the dependent variable. GDP per capita is at purchasing power parity. The additional controls include a quartic in age, interacted with sex, plus indicators for age and sex missing. We report robust standard errors, clustered at the country level, in parentheses. For further details on the standardization of satisfaction, the exact wording of satisfaction question, and the sources for GDP per capita, see the text. ***, ** and * denote statistically significant at 1%, 5% and 10%, respectively. Sacks, Stevenson & Wolfers, Income and Happiness

Comparing Within and Between Country Estimates βbetween ≈ βwithin ≈ 0.35 Sacks, Stevenson & Wolfers, Income and Happiness

Are There Satiation Effects? There exists no dataset in which the well-being-income gradient changes at incomes above $15,000 There exists no income level above which further income does not raise happiness Source: Betsey Stevenson and Justin Wolfers (2010), “Subjective Well-being and Income: Is There Any Evidence of Satiation?” Stevenson and Wolfers, Subjective Well-Being

Outline: Assessing the Happiness-Income link Within-country comparisons USA All countries Between countries: Through time Multiple datasets Both happiness and life satisfaction No evidence of satiation National Time Series and International Panels Japan Europe World Values Survey βwithin ≈ 0.35 βbetween ≈ 0.35 So let’s now turn to the national times series evidence: do societies get happier through time as they become richer. “income growth in a society does not increase happiness”. - Easterlin (1995) Sacks, Stevenson & Wolfers, Income and Happiness

Happiness & Economic Growth: Time Series Evidence Japan Old view: Life satisfaction flat, despite robust post-war growth New view: Robust growth in life satisfaction …apparent only when taking account of breaks in the data US Old view: No increase in happiness since 1972 despite GDP growth New view: True. …but few sample participants experienced income growth Europe Old view: No clear evidence of growing life satisfaction (1973-89) New view: Extending data through 2007 yields clearly rising satisfaction Around the World Old view: No clear evidence of growing life satisfaction New view: Clear positive relationship between income and growth … when comparing comparable samples Stevenson and Wolfers, Subjective Well-Being Source: Betsey Stevenson and Justin Wolfers (2010), “Time Series Evidence on the Well-Being Link”

Time Series: No rise in happiness, despite growth Japan U.S.A Europe Life in Nation Surveys General Social Survey Eurobarometer Arguably the best evidence that national happiness does not increase with economic growth came from Japan. The Japanese government had been collecting subjective well-being data since 1958 and Easterlin, as shown here, argued that Japan has had enormous growth—with GDP increasing 6-fold in the post-war period and yet there is no evidence of an increase in national well-being. Similarly, Easterlin found mixed evidence in Europe using the Eurobarometer and no increase in happiness in the US since 1972. We start with Japan, even though this one feels a bit like cheating because we did something really simple: we looked at the original codebook and hired a translator to translate the questions from Japanese to English. It turned out that there were some very important coding breaks in the Japanese data. In 1964, the response categories changed dramatically. Specifically, the top category was changed from the catch-all “Although I am not innumerably satisfied, I am generally satisfied with life now” to the more demanding “Completely satisfied”, leading the proportion reporting their well-being in this highest category to decline from 18.3% to 4.4%. (the second top category becomes somewhat more demanding, changing from: “although I can’t say that I am satisfied, if life continues in this way, it will be okay”, to “although I can’t say I am completely satisfied, I am satisfied”.) Questions asked from 1958-69 focused on feelings about “life at home”, rather than global life satisfaction (which was the focus of the relevant question from 1970 onward). The survey question—and the allowable responses—change again in 1992. Accounting for these series breaks yields a very different perspective. Sacks, Stevenson & Wolfers, Income and Happiness

Japan: Well-Being versus GDP Closer inspection reveals that the Japanese data are not as pervasive as many thought. Given the survey breaks we are left with four periods within which we can examine the change in SWB consistently. So we start with the simple approach: treating these data as four separate datasets, and following our earlier style of analysis. Doing this we see that clear evidence of a positive relationship between happiness and income until the last period during which there was little growth and large increases in unemployment. Within each continuous sub-series, we create a time series of average well-being by an ordered probit of subjective well-being on survey fixed effects. By construction, the levels of these series are not comparable, and hence comparisons within, but not between series are valid. This figure shows the satisfaction-GDP gradient within each of these periods, and it is clear that throughout the period in which Japan moved from poor to affluent (shown in the first three panels), subjective well-being rose with per capita GDP. The right-most panel shows that since 1992 the Japanese economy has shown very little growth, and subjective well-being has fallen sharply. βtime series ≈ 0.35 Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being Japan: Raw data We can put these series together and compare them to the changes in GDP and unemployment that Japan has experienced. The top panel shows real GDP and unemployment and points to roughly three episodes in Japanese economic history, corresponding roughly to changes in the well-being question asked: spectacular growth during the period 1958-69 spanning one series break in the well-being question; slower growth from 1970-91; and then anemic growth from 1992 onward, which coincided with the emergence of large-scale unemployment. The dotted markers in the bottom panel of the figure show the corresponding (and non-comparable) movements in subjective well-being within each of the periods for which consistent data exist. The bottom panel shows the raw series with breaks and then the series pooled using regression analysis to account for the series breaks. Stevenson and Wolfers, Subjective Well-Being

USA: Is it surprising that happiness hasn’t grown? We have computed the rise in income inequality in both the CPS and GSS samples. From 1972-2006, the CPS measure the log of average real household income rose by 41 log points, while inequality—measured by the mean log deviation—rose by 19 log points. Together, these numbers imply that the average level of log household income rose by only 22 log points over the full sample. For the GSS, the rise in the log of average family income is slightly smaller, at 32 log points, while the measured rise in inequality (again measured as the mean log deviation) is 15 log points. Thus, within the GSS sample, the average level of the log of family income has risen by only around 17 log points since 1972 (an annual rate of growth of only around 0.5% per year). Based on a happiness-income gradient of around 0.2-0.4, it seems reasonable to expect happiness in the United States to have been basically flat over the past 35 years. Thus, by re-focusing our attention on the appropriate macroeconomic aggregate (in the bottom panel of Figure 18), it can be seen that the U.S. experience is still roughly consistent with the evidence of a robust happiness-income link. It is worth being a bit more explicit about how reasonably robust economic growth translates into weaker growth in the average log income. From 1972-2006, real GDP per capita grew by 93%, or 66 log points, and disposable personal income per capita rose by a similar amount. Beyond these aggregate BEA data, the Census Bureau also calculates income per household from the March CPS. These alternative data suggest that personal income per capita (in 2005 dollars) rose by 65.8%, or that the log of personal income income rose by 51 log points. Much of the gap between the BEA and CPS measures reflects differences in deflators. (From 1972-2006 the CPI-U-RS (used by the Census Bureau to deflate the CPS data) rose 11 log points more than the GDP deflator. This difference would be even larger (22 log points) were we to deflate instead by the official CPI-U series.) On a per household basis, the rise in the log of average income was even less impressive, at only 41 log points. We have deflated the GSS income data using the CPI-U-RS, rather than the CPI-U-X. If instead we used the official deflator, we would register barely any growth at all in the average log of family income. Happinesst = 0.048 * Average log household income in GSSt [95% ci: -0.25 - +0.34] Happinesst = 0.058 * Average log household income in CPSt [95% ci: -0.21 – 0.33] Sacks, Stevenson & Wolfers, Income and Happiness

Change in Life Satisfaction and Economic Growth in Europe Easterlin (1995, p.38) also analyzed these nine countries (through to 1989): “Satisfaction drifts upward in some countries, downward in others. The overall pattern, however, is clearly one of little or no trend in a period when real GDP per capita rises in all of these countries from 25 to 50 percent.” This analysis is similar to that in Easterlin 1995, the solid dots show the data that he analyzed. With the benefit of time we have been able to add the observations represented by the hollow dots (an additional 18 years of data). We believe that the increased data has increased support for an increase in life satisfaction with rising income in Europe. 8 of the 9 countries have rising SWB and 6 are statistically significant. However, our point is not to count up the number of statistically significant responses one way or the other, but rather to suggest that across these nine large European countries, life satisfaction has typically risen with GDP. Moreover, estimates of the satisfaction-GDP gradient based on these national time series, while quite variable, are centered around 0.2, with some being larger, and some smaller. The upward trend in life satisfaction is not well understood and we suspect this has to do with the addition of poorer and less happy countries to the EU and hence Eurobarometer—the next figure illustrates this point. βtime series ≈ 0.35 Stevenson and Wolfers, Subjective Well-Being

Change in Life Satisfaction & Economic Growth: World Values Survey βtime series ≈ 0.35 Stevenson & Wolfers, Economic Growth and Happiness

International Panel Data: World Values Survey Sacks, Stevenson & Wolfers, Income and Happiness

Satisfaction and Growth: International Panel Regressions Sacks, Stevenson & Wolfers, Income and Happiness

International Panel Data: Eurobarometer Sacks, Stevenson & Wolfers, Income and Happiness

Long Differences: World Values Survey βpanel ≈ 0.35 Sacks, Stevenson & Wolfers, Income and Happiness

Long Differences: Eurobarometer βpanel ≈ 0.35 Sacks, Stevenson & Wolfers, Income and Happiness

Conclusion: Stylized Facts about Well-being and Income Within-country comparisons USA All countries Between countries: Since 1946 Multiple datasets Both happiness and life satisfaction No evidence of satiation National Time Series Japan 9 European nations International Panel data World Values Survey Eurobarometer Other measures of well-being Pain, depression, smiling, good tasting food to eat, respect βwithin ≈ 0.35 βbetween ≈ 0.35 βtime series ≈ 0.35 Let me turn to open thread, before moving on to discussing other measures of well-being Satiation βinternational panel≈ 0.35 Sacks, Stevenson & Wolfers, Income and Happiness

Is there any evidence of satiation? “if we compare countries, there is no evidence that richer countries are happier than poorer ones – so long as we confine ourselves to countries with incomes over $15,000 per head.” - Layard (2005) Rich countries (GDP>$15,000) Satisfaction=1.08*log(GDP) [se=0.21] Poor countries (GDP<$15,000) Satisfaction=0.35*log(GDP) [se=0.04] A 1% rise in GDP: Has three times larger effects in rich countries than poor countries A $100 rise in GDP 3x larger effect in Jamaica than US 20x larger effect in Burundi than US A rise in income of $100 is associated with a rise in well-being for poor countries that is about twice as large as for rich countries 1% rise in US GDP ~10% rise in Jamaican GDP [Comment on “20x larger”] 1/60th Test for satiation, running the regressions separately for countries w/ incomes below $15,000. Point estimates are, on average, 3 times larger for countries w. incomes averaging $15,000 One explanation for why previous authors may have thought there was a satiation point is b/c of linear scale does flatten out, but slope doesn’t go to -. In a levels specification the SWB-income gradient is curvilinear + thus less steep among wealthier countries. Sacks, Stevenson and Wolfers, Subjective Well-Being

Is there any evidence of satiation? “if we compare countries, there is no evidence that richer countries are happier than poorer ones – so long as we confine ourselves to countries with incomes over $15,000 per head.” - Layard (2005) Rich countries (GDP>$15,000) Satisfaction=1.08*log(GDP) [se=0.21] Poor countries (GDP<$15,000) Satisfaction=0.35*log(GDP) [se=0.04] A 1% rise in GDP: Has three times larger effects in rich countries than poor countries A $100 rise in GDP 3x larger effect in Jamaica than US 20x larger effect in Burundi than US A rise in income of $100 is associated with a rise in well-being for poor countries that is about twice as large as for rich countries Sacks, Stevenson & Wolfers, Income and Happiness

Happiness and GDP (World Values Survey) Sacks, Stevenson & Wolfers, Income and Happiness

Happiness versus Life Satisfaction To better understand whether the happiness-GDP gradient systematically differs from the satisfaction-GDP gradient, we searched for other data collections that asked respondents about both happiness and life satisfaction. Figure 6 brings together two such surveys: the 1975 Gallup-Kettering survey and the First European Quality of Life Survey, conducted in 2003. In addition, the bottom panels of figure 6 show data from the 2006 Eurobarometer, which asked about happiness in its survey 66.3 and life satisfaction in survey 66.1. In each case the happiness-GDP link appears to be roughly similar to the life satisfaction-GDP link, although perhaps, as with the World Values Survey, slightly weaker. Sacks, Stevenson & Wolfers, Income and Happiness

More Income is Associated with Less Pain More Enjoyment Less Depression BUT More Stress How do people’s days differ with income? More likely to be treated with respect More likely to eat good tasting food More likely to smile or laugh And people are more likely to want to have more days like the one that they just had Stevenson and Wolfers, Subjective Well-Being

Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness Yesterday Sacks, Stevenson & Wolfers, Income and Happiness

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being Now you might think that talking about economic development and love is a good way to close. Perhaps, but it may undermine the more serious message here. Stevenson and Wolfers, Subjective Well-Being

Stevenson and Wolfers, Subjective Well-Being Source: Alan Krueger, Betsey Stevenson and Justin Wolfers, “A World of Pain”.

Stevenson and Wolfers, Subjective Well-Being A Global Map of Pain Note: Our pain data covers 98.5% of the world’s population. Stevenson and Wolfers, Subjective Well-Being Source: Alan Krueger, Betsey Stevenson and Justin Wolfers, “A World of Pain”.

Sacks, Stevenson & Wolfers, Income and Happiness Blank slide: End of talk Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness Research Program Research program (with Betsey Stevenson) that seeks to establish robust stylized facts about subjective well-being Relationship between income and happiness “Economic Growth and Happiness: Reassessing the Easterlin Paradox” Consequences of income inequality “The Happiness Price of Income Inequality” (in progress) Happiness and gender “The Paradox of Declining Female Happiness” Subjective well-being and the business cycle “Is Business Cycle Volatility Costly?” Trends in the distribution of happiness “Happiness Inequality in the United States” Affect versus evaluative measures of well-being “A World of Pain” (also with Alan Krueger) NOT: asking what happiness means ARE: challenging conventional wisdom about happiness ARE: establishing foundational facts We all know the facts about what has happened to income inequality in the U.S. (rising sharply) This paper: New fact: happiness inequality is declining Given that we will be juxtaposing these two facts: it is useful to review what we know about the link between income and happiness Sacks, Stevenson & Wolfers, Income and Happiness

Subjective Well-Being and the Business Cycle

Subjective Well-Being and the Business Cycle Stevenson and Wolfers, Subjective Well-Being

New Opportunity: Daily Data Stevenson and Wolfers, Subjective Well-Being

The Paradox of Declining Female Happiness In 1974, Richard Easterlin wrote a paper titled “Does Economic Growth Improve the Human Lot? Some Empirical Evidence”. This paper launched a literature of economists using subjective well-being data to assess the policy The first part of the title “Does Economic Growth improve the Human Lot” points to his ambitions The second part, “some empirical evidence” hints at his approach: He analyzes data from surveys asking people how happy they are, and this paper launched the economic analysis of happiness data. In recent decades, that literature has greatly expanded to hundreds of papers The “Easterlin Paradox” is easily the most cited finding in the economics of happiness, and the 1974 paper also launched a thirty-year research program for Richard Easterlin exploring the link between income and happiness. FLIP TO NEXT SLIDE HERE So, what is the Easterlin Paradox? 3 comparisons of the link between income happiness appear to give different answers Between rich and poor people in a society at a point in time: Strong evidence that income matters As far as I am aware, in every representative national survey ever done a significant bivariate relationship between happiness and income has been found.” - Easterlin 1974 Between rich and poor countries: No evidence, or weak evidence that income matters the happiness differences between rich and poor countries that one might expect on the basis of the within country differences by economic status are not borne out by the international data.” – Easterlin 1974 Within countries through time: Strong claim that economic growth does not change happiness “income growth in a society does not increase happiness” – Easterlin 1995 Title of Easterlin 1995: Will Raisng the Incomes of All Increase the Happiness of All? The juxtaposition of these three findings is the Easterlin Paradox Our task in this paper – joint with Justin Wolfers– is to revisit and revise the stylized facts about the links between income and happiness Our finding: Much of what we thought we knew, we don’t. Betsey Stevenson and Justin Wolfers Wharton School, University of Pennsylvania and NBER UC Berkeley Psychology and Economics Seminar, October 14 2008.

Happiness in the United States Sacks, Stevenson & Wolfers, Income and Happiness

Inequality and Subjective Well-Being Talk runs 12:35-1:20pm. Joint work with Betsey Stevenson—very early work in progress Betsey Stevenson Wharton School, University of Pennsylvania; CESifo and NBER Justin Wolfers Wharton School, University of Pennsylvania; CEPR, CESifo, IZA and NBER

Stevenson & Wolfers, Inequality & Happiness Research Question Does inequality undermine average levels of subjective well-being? Surprisingly mixed findings: “The presumed link between equality and happiness fails to appear, at least where income inequality is concerned. Average happiness is as high in countries with great income inequality as in nations where income differences are small” –Veenhoven (2000) Including inequality in our regressions using the available data, we find that is has a negative effect, although its’ size and significance levels are both low. –Di Tella and MacCulloch (2008) “In the whole World Values Survey sample, the Gini coefficient has a positive (albeit not always significant) effect on life satisfaction.”—Guriev and Zhuravskaya (JEP 2008, p.157) “Adding a World Bank estimate of the Gini coefficient for each national economy added no explanatory power to the well-being equation.”—Helliwell (2003) Stevenson & Wolfers, Inequality & Happiness

A Simple Framework Individual well-being: Average well-being in each country: Which implies: Average of log income, not Log of average income Equal and opposite effects Stevenson & Wolfers, Inequality & Happiness

Conditional Relationship Between GDP & MLD Stevenson & Wolfers, Inequality & Happiness Source: Betsey Stevenson and Justin Wolfers, “Inequality and Subjective Well-Being”.

Distribution of Log GDP and Inequalty Stevenson & Wolfers, Inequality & Happiness Source: Betsey Stevenson and Justin Wolfers, “Inequality and Subjective Well-Being”.

Happiness Inequality in the United States Journal of Legal Studies, June 2008 Betsey Stevenson and Justin Wolfers Wharton School, University of Pennsylvania and NBER Wharton August 20, 2008

Aggregate Happiness Trends: Comparing Dist’ns This chart shows the trends in the dispersion and mean of happiness, based on alternative functional form assumptions It turns out that our findings are invariant to sensible functional forms. This makes sense given the earlier finding that the only interesting thing happening is that %pretty happy is rising: this points to declining inequality pretty much however you measure it. More fully: We assess the robustness of our estimates of the time series of the distribution of happiness in Figure 3, based on three increasingly fat-tailed assumptions about the distribution of the latent happiness variable: normality, a logistic distribution, and a uniform distribution. While any set of assumptions will seem arbitrary, it is worth noting that the more widely-used approach presented in Figure 2 is based on the particularly unappealing assumption that the happiness distribution has three equally-spaced mass points corresponding with the three allowable responses. The key finding is that none of the qualitative (and indeed, quantitative) implications of our earlier analysis are much changed by alternative approaches to cardinalizing the happiness question, which is quite reassuring. Again, we find only a mild negative trend in average happiness, but a clear decline in happiness inequality, with a turning point registered in about the late-1980s, and only a gradual increase in the subsequent years. By any measure, happiness inequality in the first one-third of our sample period is higher than in the final third. Figure 1 provides some simple intuition for why alternative distributional assumptions yield such similar results: the decline in inequality through to the late 1980s is roughly equally evident whether looking at those who are unusually happy, or when looking at those who are unusually unhappy, and as such, placing different weights on the proportion “very happy” relative to the proportion “not too happy” yields similar trends. Sacks, Stevenson & Wolfers, Income and Happiness

Sacks, Stevenson & Wolfers, Income and Happiness Estimating this full regression yields 26 separate year fixed effects for both the level and dispersion of happiness, for each of seventeen different demographic groups (four education groups, two genders, two racial groups, two marital statuses, three age groups and four regions), for a total of 26*2*17=884 coefficient estimates. Thus, instead of showing a regression table, we present these point estimates graphically in Figure 6 (focusing on average happiness levels for each group Rather than showing coefficients relative to an arbitrary omitted group, each panel shows the predicted levels and dispersion of happiness of someone with the average sample characteristics, except for the particular characteristic examined in each panel. Thus, for instance, the top (thick) line in the first panel of Figure 6 shows the evolution of happiness for someone with college education, but all other (non-education) covariates set to their (time-invariant) sample average (and Figure 7 shows the corresponding variance). These figures illustrate in more detail the broad trends seen in Table 3. Figure 6 shows that happiness has fanned out by education, with happiness highest (and rising) for college graduates, but lower and falling for high school graduates, and declining more steeply for high school dropouts. Happiness has also become more disperse by age, albeit only slightly. Among 35-49 year olds and those over 50, happiness has trended downward, while the happiness of 18-34 year olds has trended upward. As previously seen, happiness has converged along gender and racial lines. Indeed, the closing of the racial happiness gap is striking and appears to have nearly been eliminated in recent years, suggesting that the much larger unconditional racial happiness gap seen in Table 2 may be attributable to the combined impact of racial differences in educational attainment and widening educational differences in happiness. Again we emphasize differences by demographic group are merely descriptive means (or conditional means) and it should not be inferred that these are causal relationships. Sacks, Stevenson & Wolfers, Income and Happiness