The Easterlin Paradox: Empirics on the Income- Happiness Relationship sans Hedonic Adaptation Dr. Edsel L. Beja Jr. ATENEO DE MANILA UNIVERSITY
The Easterlin Paradox: Empirics on the Income-Happiness Relationship sans Hedonic Adaptation
Dr. Edsel L. Beja Jr.
ATENEO DE MANILA UNIVERSITY
Easterlin Paradox
Easterlin
• long run relationship between income and happiness is nil
Stevenson-Wolfers
• long run relationship between income and happiness is positive (and stat. significant)
Methodology: Intro
• hi = b0 + b1 gi + errori,
where:
hi = ΔH, and H is ave happiness of country
gi = ΔY, and Y is log income of country
• using h as dependent variable implies no happiness adaptation
Methodology: Intro
• H = f(Y – AY) where AY is adaptation level
• ΔH = f(ΔY – ΔAY)
• Define AY = aY-1 + (1 – a)AY-1 … where a represents the rate of adaptation
• Rearranging, ΔAY = a(Y-1 – AY-1)…
• Substituting… ΔH = ΔY – a(Y-1 – AY-1)
• ΔH = f(ΔY – aH-1)
• Or, H = ΔY + (1 – a)H-1
Methodology: Intro
• ΔH = f(ΔY)
• hi = b0 + b1 gi + errori,
where:
hi = ΔH, and H is ave happiness of country
gi = ΔY, and Y is log income of country
Methodology: Dynamic
• hit = a0 + Σbj gi,t-j + Σdk-1 hi,t-k + errorit
• where:
j = 0…p lags
k = 1…q lags
• autoregressive distributed lag model with p lags on economic growth and q lags on happiness on the assumption that current and past information on both economic growth and happiness are relevant.
Methodology: Multilevel
• hit = a0 + Σbj gi,t-j + Σdk-1 hi,t-k + errorit
a0 = φ00 + u0t
• where:
φ00 is b/w country-averages across time
u0t is the b/w country-averages variation
• hit = φ00 + Σbj gi,t-j + Σdk-1 hi,t-k + (u0t + errorit)
Results
Dynamic panel Multilevel reg
Growth(t) 0.002, p = 0.095 0.001, p = n.s.
Growth(t-1) 0.003, p < 0.001 0.003, p < 0.05
• Sum (dynamic) = 0.005, p < 0.01 0.0038 LR
• Sum (multilevel) = 0.004, p < 0.05 0.0032 LR
Analysis
• A unit of (LR) economic growth has 0.003 impact on happiness
• If ave. long-run growth rate is 1.95, then 0.003 x 1.95 = 0.00585
• Suppose ave. happiness is 3.0 in a 4 unit scale.
• To raise ave. happiness to 3.1, then 0.1 / 0.00585 = 17.09 years.
Conclusion
• Reject the Easterlin Paradox based on the statistical significance of results (i.e. that is, there is indeed positive income-happiness relationship)
• Accept the Easterlin Paradox based on the economic non-significance of results. The estimates can be taken as a refutation of the Easterlin Paradox if they have econo-mic meaning.