How does monetary policy affect welfare? Some new estimates using data on life evaluation and emotional well-being

Models on the optimal design of monetary policy typically rely on a welfare loss function defined over unemployment and inflation. We estimate such a function using two different dimensions of well-being. The first evaluates how close one is to ‘the best possible life’ on a ladder scale. The second captures the emotional quality of everyday experiences. Our Gallup World Poll sample covers 1.5 million people in 141 nations from 2005-2019. Unemployment and inflation reduce well-being across all measures. The ratio of the unemployment-to-inflation effect is 6.2 for the ‘ladder-of-life’. It is lower for positive day-to-day experiences and higher for negative ones. ______________


I. INTRODUCTION
Many Central Banks have enacted legislation that ties their ultimate purpose to promoting the welfare or well-being of their countries. 1For example, the Reserve Bank of Australia Act (1959)   states that "it is the duty of the Reserve Bank Board ... to ensure that the monetary and banking policy of the Bank is directed to … the economic prosperity and welfare of the people of Australia". 2 In addition, whilst the primary objective of the European Central Bank is to maintain price stability, its broader aim is to support "the achievement of the objectives of the European Union" which includes promoting "the wellbeing of its peoples". 3In another context, the Central Bank of the Russian Federation links its conduct of monetary policy to the purpose of achieving "stable well-being of Russian people". 4rthermore, the Reserve Bank of New Zealand Monetary Policy Amendment Act (2018) gives it a (new) aim of promoting "the prosperity and well-being of New Zealanders …" via the implementation of monetary policy directed at "achieving and maintaining stability in the general level of prices … and supporting maximum sustainable employment".
The present paper addresses the question of how to estimate a social welfare function defined over inflation and unemployment that can help inform the policy decisions of central banks by providing a way to measure the well-being costs arising from macroeconomic fluctuations.
Consider the case of a Central Bank seeking to achieve price stability in an economy which currently has a high inflation rate.It faces a decision regarding how much to increase interest rates.The more it increases rates, the more unemployment may be pushed up and the quicker inflation may be reduced, at least in the short-run, according to the Phillip's curve trade-off. 1See Mishkin (2007).
each respondent to evaluate their life on a scale in which 0 is "the worst possible life" and 10 is "the best possible life".This question is regarded as capturing the thoughts that people have about their whole life when they think about it.A related survey question with a similar type of broad focus asks individuals to evaluate whether, "on the whole", they are "satisfied with the life they lead", which was used by Blanchflower et al (2014).
On the other hand, there is another dimension of well-being that relates more to the emotional quality of an individual's everyday experience that has been distinguished by recent research.Deaton and Kahneman (2010) find that life evaluation rises steadily with log income whereas there is no further progress in emotional well-being beyond an annual income of US $75,000.
The emotions include the frequency and intensity of a person's experiences of joy and anger that make one's day-to-day life either pleasant or unpleasant. 5This dimension is regarded as being associated with instantaneous, or momentary, levels of well-being and has a short-term focus.By contrast, the evaluation-of-life questions referred to above are regarded as having a longer time horizon.Consequently, we raise the question of whether inflation and unemployment affect wellbeing differently across these different dimensions. 5These pleasant and unpleasant emotions are sometimes referred to as "positive" and "negative" affect when constructing measures of well-being (see Myers and Diener, 1994).
The paper is organized as follows.In the next section we show how the literature on the costs of macroeconomic fluctuations can be used to estimate a welfare function defined over inflation and unemployment.In section III survey data on well-being from the Gallup World Poll is correlated with these variables.The coefficients can be used to determine the welfare costs of unemployment compared to inflation.This exercise yields a different set of estimates to those often used by economists who analyze monetary policy.Section IV explains how our results may affect the design of an optimal disinflationary path and discusses the question of which measure of well-being should be used.Section V concludes.

II. THE COSTS OF INFLATION AND UNEMPLOYMENT: SOME THEORY
Central banks often emphasize many different types of costs of inflation in their publications.
First, inflation may induce people to spend additional time and mental energy on reducing their holdings of cash, rather than on more productive activities.Second, it can cause firms to incur greater 'menu costs'.Third, since tax laws are mostly not indexed, inflation may raise effective tax rates and reduce economic growth.Fourth, it can make economic calculations harder since the currency becomes less reliable as a yardstick for measuring value.Fifth, because unexpected changes in prices redistribute real wealth between debtors and creditors, volatile inflation may create risks, making the use of long-term contracts using money as the unit of account less tenable.Sixth, when price adjustments are staggered, inflation can introduce spurious volatility in some prices relative to others, reducing a market system's ability to allocate resources efficiently. 6ese different channels vary in terms of importance.For example, since holdings of cash are 6 Fischer and Modigliani (1978) was one of the first papers to outline the different costs of inflation.usually quite small, they are unlikely to justify the observed focus on keeping inflation low. 7stead, attempts to derive high costs of inflation have been more influential when focusing on the extent to which inflation reduces the price system's ability to allocate resources efficiently. 8 In addition to the above "standard costs" of inflation, other types of costs stemming from the psychological processes of individuals have also been described by behavioural economics.As an example, Shiller (1997) shows that when asked direct questions about inflation, people report 'unconventional' problems, like exploitation, lower national prestige and a loss of morale.One reason may be that the fairness of the existing income distribution becomes harder to justify when there is confusion over prices due to high inflation.For example, if speculation is more common when relative price changes occur more frequently, then one may find it harder to claim that the economy is rewarding effort more than luck.People may also experience regret that they didn't buy when prices were lower. 9 With respect to the costs of unemployment, in spite of a long tradition studying macro-economic fluctuations, there remains disagreement among economists about the seriousness of their effects.The welfare costs of recessions in classical economics arise from the lost output that occurs when actual output falls below potential.This approach is sometimes adopted by realbusiness-cycle theorists, who assume that individuals are optimizing and recessions are desirable adjustments to productivity shocks.This means that the costs of business cycles are smallperhaps only 0.1 percent of total consumption in the US. 10 7 See Friedman (1969). 8 This is the approach taken by Rotemberg and Woodford (1997). 9See Rotemberg (2005, 2009). 10See Lucas (2003).
By contrast, and along similar lines to better understanding the costs of inflation, there may also be potential for behavioural economics to improve our knowledge of the different kinds of costs associated with recessions.Substantial work in psychology and sociology indicates that there are emotional costs to those who lose their jobs far exceeding the monetary costs. 11For example, the unemployed may experience a loss of social status and self-esteem.They can find themselves deprived of personal relationships which they gained through their job and lose the structure that comes through being bound to a workplace.In addition, the unemployed may suffer a stigma due to being labelled as lazy, as well as face blame for their situation.

III. ESTIMATING A SOCIAL WELFARE FUNCTION
When it comes to formulating monetary policy, knowing the welfare effects on a society of both inflation and unemployment is of first order importance.However due to the potentially many different psychological costs that may occur when these variables change, as outlined above, identifying the total cost is difficult.We remain a long way from having useful estimates to help guide policy-makers.The direct questions about the costs of inflation that were used by Shiller   (1997) are subject to notable criticisms.Diamond and Hausman (1994), for example, argue that there can be strategic manipulation of answers in contingent valuation studies of environmental costs which use a similar style of question.Much may also depend on the respondent's ability to understand difficult issues (such as the workings of the economy or state of the environment).
Another approach is to ask people about their own personal level of well-being and then correlate the answers with our variables of interest (i.e., inflation and unemployment).This imposes fewer informational demands, as presumably it is easier to know about one's own situation, than about how the economy works. 12In other words, our approach is to estimate a welfare function of the following form: The task of theoretically deriving a function of this type was undertaken by Rotemberg and   Woodford (1997) who ground their structural relations in the context of optimizing behavior of individuals and of firms that must temporarily keep their prices fixed, resulting in relative price distortions when inflation rises. 13Their key assumptions are a) a summary measure of utility exists; b) all channels through which inflation and unemployment matter can be reduced to consumption and leisure; c) there is a representative agent.
The advantage of determining welfare losses arising from changes in inflation and unemployment in the same model is that one can then directly compare their relative size, which is important for Central Banks when seeking to reduce inflation by raising interest rates since it helps to pin down the optimal adjustment path.14Specifically, we estimate the regression: The response categories are shown to individuals as options from which to choose.The bottom category ("0") is labelled as being the "Worst possible" life whereas the top category ("10") is labelled the "Best possible" life.In addition to these survey data, Inflation is measured by the rate of change in the Consumer Price Index and Unemployment is the number of people without work but available for (and seeking) employment as a proportion of the total labor force. 16r data set is comprised of repeated cross-sections of around 1,000 individuals living in 141 countries between 2005 and 2019, making up a full sample size of 1,489,290 individuals.The Appendix reports the names of these countries, as well as summary statistics.Although the Gallup World Poll has well-being data on 168 countries, this number reduces to 141 countries for which both unemployment and inflation rate data are also available. 17 Whereas the ladder-of-life question has 11 response categories, the questions on daily experiences like sadness and enjoyment that we use in the next section have only 2 response categories.The World Poll mostly uses "a simple dichotomous (yes or no) response to minimize contamination of data because of cultural differences in response styles and to facilitate cross-cultural comparisons". 18We follow a similar methodology to Deaton and Kahnemann (2010) who also use these kinds of data (in their case, to determine how income affects well-being in the US).The coefficients on inflation and unemployment estimated by equation ( 2) provide a way to aggregate all the costs and benefits of macro-economic fluctuations.In other words, the regression patterns map out a welfare function, unbeknown to the respondents completing their well-being survey score sheets. 19ble I presents simple Ordinary Least Squares results when the 11 point Ladder of Life scale is regressed on inflation and unemployment.Column (1) controls for country and year fixed effects.
The coefficients on the unemployment and inflation rates are both negative and significant at the 1 percent level.In order to see the size of the effects, a 10 percentage point increase in the unemployment rate implies ticking down one's well-being score by 0.59 points on the 0 to 10 17 The 27 countries for which either unemployment or inflation data, or both, are unavailable are Gabon, Guinea, Guyana, Cuba, Lesotho, Maldives, Oman, Puerto Rico, Somalia, Eswatini, Gambia, Kazakhstan, Belize, Bosnia & Herzegoniva, Central African Republic, Comoros, Congo (Kinshasa), Djibouti, Palestinian Territories, Uzbekistan, Turkmenistan, Argentina, Taiwan, Somaliland, Northern Cyprus, Nagorno-Karabakh Region and South Sudan. 18See the Gallup Worldwide Research Methodology and Codebook at https://data-services.hosting.nyu.edu/wpcontent/uploads/2017/10/World_Poll_Methodology_102717.pdf.A dichotomous dependent variable also avoids issues associated with using ordered probit regressions that are run on well-being response data with three or more categories (see Bond and Lang, 2019). 19To validate well-being data, researchers have established a connection between them and objectively measured variables that are associated with 'true utility'.For example, studies reveal that unemployed individuals report low levels of well-being, even after controlling for the income drop associated with job loss.This association is important since objective factors like addiction, depression and violence (that may negatively affect well-being) are often linked to unemployment.There is also a positive connection between well-being survey scores and (observed) good health outcomes (e.g., Stutzer and LaLive, 2004, Blanchflower and Oswald, 2008).Furthermore, well-being responses are correlated with left frontal brain activity, which is connected to positive emotional states.
scale (=5.86*0.10). 20Meanwhile, an increase in the inflation rate of 10 percentage points reduces one's well-being score by 0.11 points (=1.10*0.10).The size of the ratio of the effect of unemployment to inflation is 5.3 (=0.586/0.110). 21 <Insert Table I here> It could be argued that the above calculation underestimates the cost of unemployment since there are two unpleasant consequences of a rise in unemployment: some people lose their jobs while at the same time everyone in the economy becomes more fearful.Using the coefficient on the unemployment rate in column (2) we can calculate that an increase in the unemployment rate of one percentage point has a cost in well-being units equal to 0.0646 for the average citizen (=6.46*0.01).This number may be viewed as capturing the "fear of unemployment" effect that occurs across all workers.Meanwhile, an individual who actually falls 20 For cross-sectional evidence on the relation between the "ladder-of-life" and economic fluctuations, see Gandelman and Hernández-Murillo (2009).
21 Svensson (2002) converts these estimates into a trade-off between the output gap and inflation using Okun's Law.
He states "a simple version of Okun's Law is that a change of the unemployment rate of one percentage point corresponds to a change of the output gap of 2 to 2.5 percentage points".Hence a one percentage point reduction in the output gap causes between 2.1 (=5.3/2.5) and 2.6 (=5.3/2) times as much of a reduction in well-being as an additional percentage point of inflation. 22By contrast, as one goes up the income quintiles there is a monotonically increasing effect.A rise from the bottom to top quintile increases the ladder-of-life by 1.1 units.
unemployed experiences a much larger personal cost, equal to 0.54 units.Consequently the overall well-being cost to society of a 1 percentage point increase in the unemployment rate is the sum of these two components, or 0.070 (=0.0646+0.01*0.54).The second component is the 0.54 unit cost to an individual of being unemployed multiplied by the 1 percent of the population who have been unlucky enough to actually lose their job.Taking both components into account, the ratio of the effect of unemployment to inflation rises to 6.2 (=0.070/(0.01*1.12)).As a check on robustness, columns (3-4) estimate the same regressions but instead using ordered probit.They again find that high rates of both unemployment and inflation reduce the likelihood of feeling close to the "best possible life", at the one per cent level of significance.The ratio of the size of these effects is not significantly different to the ones previously calculated. 23I.b.Estimating a Social Welfare Function using Data on Day-to-Day Feelings Whilst the "ladder-of-life" question refers to the judgments people make when they think about their overall life, psychologists often distinguish these evaluations from the day-to-day feelings that they encounter as they live it. 24The latter refer to the emotional quality of an individual's everyday experience -such as the joy or sadness that make one's life either pleasant or unpleasant. 25A natural question, given such a multiplicity of emotions, is whether measures appropriate for empirical analyses are available to produce more comprehensive tests.The World Poll, which was designed with the assistance of psychologists, does indeed collect data on this 23 A related measure of well-being is captured by the question: "On the whole, are you satisfied with the life you lead?"Most surveys provide four possible answers, such as 'not at all satisfied', 'not very satisfied', 'fairly satisfied' and 'very satisfied'.Blanchflower et al (2014) use Euro-barometer Survey Series data which asks this question and find that a one percentage point rise in the unemployment rate lowers life satisfaction by five times as much as a one percentage point rise in the inflation rate (see also Di Tella et al, 2001, 2003, Di Tella and MacCulloch, 2009, and Wolfers, 2003). 24 Kahneman and Krueger (2006) argue that life evaluation measures are best viewed as "a global retrospective judgment, which in most cases is constructed only when asked and is determined in part by the respondent's current mood and memory, and by the immediate context."They recount the dime experiment of Schwarz (1987), whereby subjects "accidentally" find a dime before filling out a questionnaire.The lucky half of the sample reports higher levels of life satisfaction. 25See Hager and Ekman (1983).
second aspect of well-being.Consequently, it allows us to investigate whether macro-economic fluctuations influence people's day-to-day feelings (in addition to their life evaluations). 26ecifically, we use two measures of both "negative affect" and "positive affect" to capture selfreported well-being under this shorter time horizon.With respect to negative affect, our first measure is the answer to the question, "Did you experience the following feelings during a lot of the day yesterday?How about sadness?", and the second is the answer to "Did you experience the following feelings during a lot of the day yesterday?How about physical pain?"In each of these cases we define a dichotomous dependent variable based on the either "yes" or "no" response of each individual.
Columns (1-2) of Table II present the "sadness" results.We run probit regressions and report marginal probabilities.Higher unemployment and inflation rates both increase the chances of feeling this negative emotion, at the 1 percent level of significance.In column (2) which controls for personal characteristics, as well as country and year effects, a rise in the unemployment rate has 9.0 times as large an effect on sadness as the same increase in the inflation rate (=(0.65+0.099)/0.083).Meanwhile, columns (3-4) show how macroeconomic fluctuations affect the extent to which people experience "pain".The impacts of unemployment and inflation are again both positive and significant at the one percent level, across both specifications.The ratio of the size of the effects in column (4), taking into account both the personal costs of unemployment as well as the fear effect experienced by everyone, is equal to 13.0 (=(0.39+0.001)/0.030).

<Insert Table II here>
With respect to "positive affect" we also use two measures.The first is the answer to the question, "Did you experience the following feelings during a lot of the day yesterday?How about enjoyment?", and the second to "Now, please think about yesterday, from the morning until the end of the day.Think about where you were, what you were doing, who you were with, and how you felt … Did you smile or laugh a lot yesterday?"Columns (1-2) of Table III present the "enjoyment" results.Higher unemployment and inflation rates both reduce the chances of experiencing this feeling, at the 1 percent level of significance, controlling for country and year effects, as well as personal characteristics.Using the coefficients in column (2), the relative size of the total effect to society of unemployment compared to inflation on enjoyment is equal to 5.0 times (=(0.35+0.065)/0.083).

<Insert Table III here>
Finally, columns (3-4) show how the unemployment and inflation rates affect the extent to which people "smile or laugh a lot".Column (3) shows that both of these variables reduce this measure of well-being at the 1 percent level.In column (4) the ratio of the total effect of unemployment compared to the inflation is equal to 3.8 times (=(0.18+0.055)/0.062).

Lagged Dependent Variable
In this section, we begin by including a lagged dependent variable in our regressions.Column (1) in Panel A of Table IV is our base-line regression using the full sample with a total of 1,068 country-by-year observations and includes both country and year fixed effects.The lagged dependent variable is positive and significant.Unemployment and inflation retain their negative effects on the ladder-of-life, both at the 1 percent level.

<Insert Table IV here>
Column (2) controls for personal characteristics by averaging the residuals for each country and year from first stage regressions of the form reported in Table B in the Appendix.These countryby-year unexplained well-being components then become the dependent variable in the "secondstage" regressions reported in Table IV. 27Once personal controls are included, the number of country-by-year observations equals 852.There is relatively little auto-regression, with a lagged dependent variable coefficient of 0.33 and our well-being data continue to be strongly correlated with macroeconomic variables.The ratio of the coefficient on unemployment to inflation is not significantly different from the previous estimates (without the lagged dependent variable).
Columns (3-6) repeat the exercise for our set of negative feelings.Across all specifications higher unemployment and inflation both increase sadness, at the one per cent level, and the sizes of the ratios remain similar to before.Panel B reports the results for positive feelings.

Differences Depending on Country Characteristics
Much variation exists across the countries in the Gallup World Poll in terms of their macroeconomic characteristics.For example, inflation hit rates of over 50% in Belarus, Venezuela and South Sudan over the sample period.At the other extreme, Chad experienced deflation of 9% per annum.Consequently, we check whether our macroeconomic variables have differing effects on well-being depending on whether the country is experiencing high or low inflation.
Panels A (and B) in Table V show the results for the high (and low) inflation countries in the sample, defined as those places with rates higher (and lower) than the mean inflation rate.The "high" sample has a mean inflation rate of 12% whereas the "low" sample has a mean rate of 2%.The most striking feature when comparing these panels is that inflation has a negative and significant adverse effect across every measure of well-being for the high inflation countries (i.e., it decreases the "ladder-of-life" in column (1), increases negative feelings like sadness and pain in columns (2-3) and lowers positive feelings like enjoyment and smiles in columns (4-5)) whereas it has no significant effect across any of these measures for low-inflation countries.Furthermore, the difference between the size of the effects for the high and low inflation countries on, for example, the "ladder-of-life", is significant at the 5 per cent level.

<Insert Table V here>
We also divided the sample according to the level of development.Our low (high) development sample is defined as those countries with GDP per capita below (above) the sample mean of $18,841 (as measured in constant 2010 US dollars).This time it is the effect of higher rates of unemployment that differ between the two sub-samples.Across every measure of well-being, unemployment has a significantly more adverse effect on well-being in low development, compared to the high development, countries (at the 5 per cent level for pain and smiles; 10 per cent level for 'ladder-of-life', sadness and enjoyment).A possible reason is that the welfare state, which mitigates against the effects of job-loss through, for example, unemployment insurance programs, tends to be more comprehensive in the wealthier countries.
Finally, we look at whether there are regional differences.These may arise due to differing institutions, macro-economic experiences or cultural factors, as examples.Panel A of Table VI focuses on South America and Panel B on Europe.The countries which make up these regions in the world are listed in the Appendix.

<Insert Table VI here>
Columns (1-5) of Panel A show that both unemployment and inflation adversely affect wellbeing in South America across every dimension, at the one per cent level of significance (with the only exception being unemployment on 'smile').However, the results are less strong for Europe where, for example, there are no detectable effects of either of these macro-economic variables on pain.Not only does unemployment cause more pain in South America than Europe, it does so also in Africa and Asia.The strongest consistent negative effects of inflation across each of our measures of well-being occurs in South America.

IV. WELL-BEING DATA and CENTRAL BANKS: A DISCUSSION
One interpretation of the above results is that individuals find both inflation and unemployment costly.The bigger the adverse effect of unemployment, compared to inflation, has implications for the optimal disinflationary path followed by a Central Bank.For example, if inflation is high and needs to be cut over the medium to long term, then a smaller increase in interest rates may be justified in order to minimize the well-being costs coming from a slow-down in the economy and rise in unemployment in the short-run.
However, without direct measurements of well-being, there is little to guide Central Banks as to the relative welfare costs of inflation to unemployment.Instead the execution of monetary policy has mostly relied on measuring, aside from inflation, actual and potential output to calculate the 'output gap', as well as the actual and 'non-accelerating inflation rate' of unemployment.Deciding how much to adjust interest rates to maximize welfare, subject to a short-run Phillip's curve trade-off, has largely been left to the discretion of Central Bank officials. 28mbiguity over the costs of inflation and unemployment has led to strident debates regarding the wisdom of events like the "Volcker disinflation" in the US, during which inflation fell from a peak of 11% in early 1980 to 4% by the end of 1983.Meanwhile, a decline of employment and output on a scale not witnessed before in the post-World War II period also occurred, leading commentators to query whether the "Volcker experiment" was due to a "new, greater weight on the Federal Reserve's objective of price stability vis-a-vis its objective of output growth and high employment" (see Friedman, 2005).However, articles by monetary economists on whether the costs of such disinflations are justified mostly provide little guidance on the adverse mental health effects of unemployment compared to inflation.For example, while Goodfriend and King (2005) note how the output losses of Volcker's disinflation "had great effect on the lives of many individuals during the period, as we recall from discussions with friends, relatives, and neighbors", these kinds of costs have typically not been formally quantified, at least at an aggregate level.
By contrast, psychologists have long detailed the nature of distress suffered by unemployed individuals, which includes feelings of anxiety, stress and depression.These kinds of day-to-day emotional costs do not solely arise from the loss of income but stem more from the sense of worth that a job gives to most people. 29The effects of unemployment on mental health have attracted heightened interest ever since the Great Depression of the 1930s.A review paper by Eisenberg and Lazarsfeld in 1938 concluded that unemployment "tends to make people more emotionally unstable than they were previous to unemployment" (p.359).Those who lose their jobs report sleeping less due to worry, feeling strained and worthless, finding it harder to concentrate and being less able to enjoy day-to-day activities (as measured by the General Health Questionnaire in the United Kingdom). 30The psychological impact of unemployment is now widely recognized.
The Canadian Mental Health Association, for example, notes that "You can experience some of the same feelings and stresses [due to unemployment] that you would if you were seriously injured, going through a divorce, or mourning the loss of a loved one.You can go through some or all of the stages of grieving just as you would with any other major loss". 31A higher unemployment rate also reduces the well-being of those who remain in work via a more pessimistic perception of their own future unemployment prospects.Anticipation of redundancy for an employed individual has been shown to be as distressing as the experience of unemployment itself, being associated with fear and anxiety. 32 Regards the mental well-being effects of inflation, Shiller (1997) finds that 86% of people report themselves as feeling angry when they see prices rise.He comments that "Inflation, when it is substantial or shows the risk of becoming substantial, is clearly perceived as a national problem of enormous proportions.This fact is also evident in the constant attention that inflation is given in the media and in the fundamental role it plays in many political elections.News about inflation seems to have serious consequences for approval ratings of presidents and for outcomes of elections.Public-opinion polls have shown that inflation … has often been viewed as the most important national problem".In Venezuela in 2015, for example, worries about inflation became the number one concern in that country, ahead of violent crime, which was surging at the time. 33 we saw in the previous section, when it comes to comparing relative sizes, unemployment has a more adverse effect than inflation on life evaluation, as well as on all of our measures of 30 See Clark and Oswald (1994). 31See https://cmha.calgary.ab.ca/mental-health/understanding-mental-illness/unemployment/ 32 See De Witte (1999) and Clark, Knabe and Rätzel (2009).Hartley et al. (1991) report how those with job insecurity have impaired mental health, as expressed by psychosomatic symptoms and depression. 33The survey was done by Caracus pollster, Datanalisis: www.bbc.com/news/world-latin-america-32703081.Inflation reached 181% per annum in Venezuela in 2015.
positive and negative experiences, though the size of the estimated ratio varies substantially across them.The reason for such variation may stem from there being a multiplicity of different emotions relevant to human experience that are not all the same. 34Moreover, positive and negative affect appear to be processed by different neural systems in which 'good' and 'bad' are independent dimensions (see Cacioppo et al., 1999).In this sense, there is no a-priori reason why variables like unemployment and inflation should have the same sized effects across our different measures.Put another way, psychologists don't view there as being a single 'correct' measure of well-being but instead argue that there are different dimensions which cannot necessarily be aggregated into a single summary measure.By contrast, economists tend to build their models in terms of a single dimension which they call "utility".Consequently, for a government that seeks to directly estimate well-being costs, whether there is one best measure that most closely captures economists' concept of utility is still open to much debate. 35ide from monetary policy, the mental well-being effects of financial crises are becoming increasingly recognized, particularly due to the actual and threatened job losses which they can cause. 36In the context of its review of the level of capital which is considered prudent for banks to hold, the Reserve Bank of NZ states "banking crises can have large and long-lasting impacts … beyond the initial economic downturn that may have precipitated them.In addition to the cost of lost output, broader societal costs of crisis events include impacts on health, mental well-being and social cohesion" (see RBNZ,   2019).More generally, in the aftermath of the Global Financial Crisis, leading commentators highlighted the greater need for "careful empirical research with attention to psychological as well as economic 34 Facial analysis is used by Hager and Ekman (1983) to code different emotions. 35For example, to the extent asset price bubbles arise from a fear of missing out, Central Banks that use life evaluation scores for policy purposes instead of more immediate feelings of 'regret' may be criticized by those who care about these phenomena.In addition, politicians who enact legislation emphasizing long-run well-being may lose office to ones who propose policies that give people positive experiences right now, at this point in time. 36 Montagnoli and Moro (2018) find that financial crises lead to "major, widespread and lasting psychological losses", above and beyond that which can be attributed to lower output.
factors". 37In addition, the World Health Organization has detailed the links between such crises and mental health, which it emphasizes largely occur through unemployment. 38Psychologists have also shown how financial crises are often associated with increases in depression and suicide rates. 39  3 See "Implications of the Financial Crisis for Economics", speech by (former) Chair of the US Federal Reserve, Ben Bernanke, at the Bendheim Center for Finance, Princeton, New Jersey, 24 September 2010. 38See World Health Organization (2011). 39Christodoulou and Christodoulou (2013).

V. CONCLUSION
This paper shows that individual subjective well-being is negatively correlated with the unemployment and inflation rates.Our contribution is to use two different types of survey questions that research has identified as capturing distinct aspects of subjective well-being.The first focusses on overall life evaluation, as measured by Cantril's "ladder-of-life", and the second on the emotional quality of an individual's (pleasant or unpleasant) day-to-day experiences, such as joy or sadness.Whereas the former has a long time horizon, the latter's focus is shorter-term.
Our Gallup World Poll sample includes 1.5 million people across 141 countries over 15 years.
In terms of the models traditionally adopted in the design of monetary policy, our estimates can be seen as a way of obtaining the weights in a social welfare function.Consequently, they can help Central Banks to understand the trade-offs that people are willing to accept in terms of unemployment for inflation.The evidence suggests that reasonable proxies for what economists call "utility" are significantly affected by aggregate-level economic fluctuations.
Our approach yields a different set of welfare weights on the unemployment and inflation rates to those often used by economists who analyze monetary policy, which typically place much less weight on the costs of unemployment.The ratio of the effect of unemployment compared to inflation equals 6.2 when Cantril's "ladder-of-life" is used as a proxy for well-being, yet rises to between 9 and 13.0 for negative day-to-day feelings and falls to between 3.8 and 5.0 in the case of positive day-to-day feelings.In summary, we believe that Central Banks should focus more of their research efforts on collecting and analyzing these kinds of data.Governments also face a challenge in terms of better defining what aspect of well-being they would like to promote.4), the number of countries drops to 136 and the years range from 2009 to 2019 (upon inclusion of personal characteristics).
[4] Cut points (standard errors) for regression (3) are -3.2 (0.1), -2.9 (0.1), -2.5 (0.1), -2.1 (0.1), -1.8 (0.1), -1.1 (0.1), -0.7 (0.1), -0.2 (0.1), 0.4 (0.1) and 0.8 (0.1) and for regression (4) are -3.1 (0.1), -2.7 (0.1), -2.4 (0.1), -2.0 (0.1), -1.6 (0.1), -0.9 (0.1), -0.5 (0.1), 0 (0.1), 0.6 (0.1) and 1.0 (0.1).    2), ( 4), ( 6), ( 8) and ( 10) are the residuals from "first stage regressions" which control for personal characteristics of the kind reported in It may have been partly due to the observed mental distress arising from high unemployment that social insurance was established in many countries in the aftermath of the Great Depression.In 1932, for example, Wisconsin became the first US state to establish an unemployment insurance program.Many public and private organizations are now measuring national well-being.One of the earliest examples is the government of Bhutan.Its' Gross National Happiness survey asks people to selfreport the frequency of experience of eleven different negative and positive emotions.Amongst developed nations, France increased measurements of subjective well-being after the report by Stiglitz, Sen and Fitoussi (2009), with the French Statistical Institute introducing questions such as "on a scale of 0 (not at all satisfied) to 10 (very satisfied), rate your satisfaction with your life at the present time".Meanwhile in the UK, questions relating to life satisfaction, happiness and anxiety gained 'National Statistics' status in 2014.Statistics Canada has fielded a satisfaction-with-life question in its annual General Social Surveys since 2002 and NZ started in 2008.The European Commission (via its Eurobarometer series which began in 1975) and the OECD (as part of its 'better life index') also collect well-being data.In the US, the National Opinion Research Centre at Chicago University has measured happiness with its' annual General Social Surveys since 1972.The World Values Survey, based in Vienna, also asks individuals to rate their own well-being.It began with 11 nations in its first "wave" (held from 1981 to 1984) and has subsequently been expanded.

Table I
How the "Ladder-of-Life" Varies with Inflation and Unemployment: 141 Countries, 2005 to 2019.Standard errors in parentheses, with country level clustering.***denotessignificance at the 1 percent level, ** at 5 percent level and * at the 10 percent level.[2] Dependent variable is response to Gallup World Poll question, "Please imagine a ladder/mountain with steps numbered from zero at the bottom to ten at the top.Suppose we say that the top of the ladder/mountain represents the best possible life for you and the bottom of the ladder/mountain represents the worst possible life for you.If the top step is ten and the bottom step is zero, on which step of the ladder/mountain do you feel you personally stand at the present time?" (An 11 point scale ranging from 0 to 10 is shown to respondents).[3]Incolumns (2) and (

Table II
How Negative Experiences vary with Inflation and Unemployment: 141 Countries, 2005 to 2019.We define Sadness to equal 1 if the response is "yes" and 0 if it is "no".Dependent variable in columns (3-4) is the individual response to the question, "Did you experience the following feelings during a lot of the day yesterday?How about physical pain?"We define Pain to equal 1 if the response is "yes" and 0 if it is "no".[3]Incolumns (2) and (4), the number of countries drops to 136 and the years range from 2009 to 2019 (upon inclusion of personal characteristics).
** denotes significance at the 1 percent level, ** at 5 percent level and * at 10 percent level.[2]Dependentvariable in columns (1-2) is the individual response to the Gallup World Poll question, "Did you experience the following feelings during a lot of the day yesterday?How about enjoyment?"WedefineEnjoyment to equal 1 if the response is "yes" and 0 if it is "no".Dependent variable in columns (3-4) is the individual response to the question, "Now, please think about yesterday, from the morning until the end of the day.Think about where you were, what you were doing, who you were with, and how you felt … Did you smile or laugh a lot yesterday?"WedefineSmile to equal 1 if the response is "yes" and 0 if it is "no".[3]Incolumns (2) and (4), the number of countries drops to 136 and the years range from 2009 to 2019 (upon inclusion of personal characteristics). *

Table IV Panel A: Further Tests
How Ladder-of-Life, as well as Negative Experiences, vary with Inflation and Unemployment: Second Stage Regressions using the "two-step method", 141 Countries, 2005 to 2019.
Notes:[1]Ordinary Least Squares Regressions are reported with standard errors in parentheses.*** denotes significance at the 1 percent level, ** at 5 percent level and * at 10 percent level.[2] The dependent variables in columns (

Table V Panel A: Further Tests
How Ladder-of-Life, as well as Negative and Positive Experiences, vary with Inflation and Unemployment: High Inflation Countries, 2005 to 2019.
*** denotes significance at the 1 percent level, ** at 5 percent level and * at 10 percent level.[2]Thedependent variables, Ladder of Life, Sadness, Pain, Enjoyment and Smile are defined in the Appendix.High inflation countries are defined as those with inflation rates above the sample mean and low inflation countries as those with inflation rates below the mean (where average inflation=0.05).

Table VI Panel A: Further Tests
How the Ladder-of-Life, as well as Negative and Positive Experiences, vary with Inflation and Unemployment: South America, 2005 to 2019.How the Ladder-of-Life, as well as Negative and Positive Experiences, vary with Inflation and Unemployment: Europe, 2005 to 2019.[1] The regression results are Ordinary Least Squares in column (1) of Panels A and B. Columns (2-5) are probit regressions with marginal effects reported.Standard errors in parentheses, with country level clustering.[2] *** denotes significance at the 1 percent level, ** at 5 percent level and * at 10 percent level.[2] The dependent variables, Ladder of Life, Sadness, Pain, Enjoyment and Smile are defined in the Appendix. Notes:

Table A
Summary Statistics: Gallup World Poll, 141 Countries, 2005 to 2019.

Table B
How Life Evaluation, Positive and Negative Feelings vary with Personal Characteristics: 136 Countries, 2009 to 2019.
Notes:[1]Column (1) is an Ordinary Least Squares Regression with standard errors in parentheses.Columns (2-5) are probit regressions with marginal probabilities reported.*** denotes significance at the 1 percent level, ** at 5 percent level and * at 10 percent level.[2] The dependent variables, Ladder of Life, Sadness, Pain, Enjoyment and Smile are defined in the Appendix.