Bulent Temel for The 2004 Moffatt Prize in Economics

GDP could be a solid measure of “national income”, but not necessarily GDP per capita is for “average national income”.  In capitalist economies in which few people get overwhelmingly rich while the majority of the nation progresses just modestly, how accurate would simple arithmetic mean be in representing average level of income?  Do the highest figures in the highly heterogeneous US income distribution cause arithmetic mean formula yield a value with a small frequency?  In other words, is US GDP per capita an income that is made by only few Americans? 

Answer to this question lies at how higher “highest incomes” are than the “others”, and what percentage of US citizens earns them.  To see where US is on the issue, let’s play with government income numbers[1] using the concept of statistical median.  Suppose we lay out all families in the United States on a giant football field.  We place them according to ascending order of their income.  (Income level rises as we move from one goal post towards the other).  Then let’s ask families make stacks of $100 bills that total their incomes.  So, for instance, the household at the 50-yard line (family with the median income) would build a stack of bills that are…

= US median income divided by dollar value of each bill, and re-divided by the height of a bill

= $40,000 / $100 / 0.004 inches

= 1.6 inches high. 

When all families erect their stacks, we end up with the below picture[2].

As seen clearly, a vast majority of American households are making annual incomes that are not even comparable with those of few wealthiest people.  The transition to this “L-Curve” from the traditional “Bell-Curve” is why GDP per capita is no longer a good measure of America’s average income.

Political philosopher David Schweickart pinpoints income inequality reality with a dramatic statement:  "…If we divided the income of the US into thirds, we find that the top ten percent of the population gets a third, the next thirty percent gets another third, and the bottom sixty percent get the last third.  If we divide the wealth of the US into thirds, we find that the top one percent own a third, the next nine percent own another third, and the bottom ninety percent claim the rest.  Actually, these percentages true a decade ago, are now out of

date.  The top one percent are now estimated to own between forty and fifty percent of the nation's wealth, more than the combined wealth of the bottom 95%.”[3]

An historical outlook on US income distribution supports the hypothesis that GDP per capita is representing real average income less and less every passing year.  If we divide all US incomes into 5 equal pieces (“quintiles”), middle quintile (Quintile-3) would be where GDP per capita –by definition- falls in.  Table-1 (Appendix) shows how percentage of 3^rd quintile in the total distribution has been declining since 1967.  In other words, GDP per capita is showing an income that is earned by fewer Americans every year.  Below chart derived from the same data, visualizes this phenomenon. 

On the other hand, not surprisingly, share of those who eat the top 5% of the pie has been growing.   

One indicator of per capita GDP’s declining validity is national progress improving slower than economic progress.  Social progress refers to the overall quality of life in a country, and is measured by “International Index of Social Progress” (IISP).  IISP, by assessing 40 aspects of life including quality of health, education, environment, level of democracy, military spending, etc; is the grade of a country’s ability to provide a good standard life.  Thanks to its multi-variate nature, IISP shows a life standard both qualitatively and quantitatively in a way superior to GDP per capita, which merely and misleadingly refers to quantitative standards.

United States, the country with the second highest average income in the world -in terms of GDP per Capita-, ranks only 27^th in social progress criterion.[5]  Richard Estes, a Social Studies Professor at University of Pennsylvania, explains why: “The failure to make progress on the wave of a historic economic boom in the 90’s is explained in part by the passage of welfare reform.  The wealth was concentrated in upper income brackets and never reached the poor.  Welfare reform meant giving the poorest less [time-limits on eligibility, work requirements with no child-care support] than what they had before.”

Top-10 countries with the highest life standards according to IISP have an unsurprising commonality: A more evenly distributed national income.  Shown by “GINI Index”, level of income inequality in these countries is lower than in other developed countries like USA and UK.  Below table that compares the levels of income inequality (GINI) and average income (GDP per Capita) in both groups helps making the point of this study:

Lower social progresses of the nations where incomes are shared unequally (USA, UK) indicate GDP per capita overestimates average incomes in these countries.  Living standards that remain relatively constant is a strong sign of GDP per Capita’s lessening meaning.    

Capitalism being criticized to “deepen the gap between rich and poor” is not a new debate.  Neither is “but today’s poor are richer” counter-argument.  Indeed, it takes only 13 minutes for an average worker to earn enough money to buy a pair of socks, a major improvement from 1.5 hour of the year 1950.  Today, people of lower income levels own more than those half a century ago.  However, merely concentrating on “how much” question, but ignoring “how much to whom” may only be an approach of hoodwinker politicians, but not socially liable economists.

A major drawback of income inequality is its contributions to economic fluctuations (which, in turn, hurts lower income groups more than others).  Combined with factors such as a weak banking system, trade deficit, contractionary fiscal policies and problematic corporate structures; income inequality leads to crises as it did in 1929 Great Depression.  In late 1920s, only 5% of Americans were earning one third of the total US income.  Top 1% owned an all-time high %36.3 of the nation’s assets.  Luxury consumptions and speculative capital investments of this society volatilized American economy.  Their excessive incomes turned back to money markets resulting in loans available to riskier borrowers.  Many people failed to pay their loans back due to their insufficient incomes.  Tax reductions in 1921, 1924, 1926 and 1928 made the situation worse by helping high-income earners piled up their disposable incomes. 

Another problem with income inequality is that it causes political instability.  When a majority of a nation is aware of their earnings being way lower than some others, they become dissatisfied with their economic status, and therefore constantly search for better governments.  Political instability increases the risks of investment and discourages foreign capital flow.  Lower investment grades of countries resulting in less funds coming from abroad; undermine a nation’s growth potential.  Income inequality also deteriorates business confidence in domestic markets.  It discourages economic entities about commitment and trust.  Higher risks of conducting business, and higher costs of enforcing contracts impede economic transactions.

Unequally shared national incomes take some economic weapons away from policy practitioners.  One of these public finance tools, “pricing” loses its functionality in highly unequal income distributions.  For instance, in needs of higher energy efficiency, unequally distributed national incomes would turn off governments to raise prices due to fear of poverty. 

“WEIGHTED AVERAGE GDP”

All above are how bad US income distribution is and the logic behind GDP per capita’s sophistry.  If per capita GDP does not work, then what does?  Stratification is where the answer is.  Earlier, we divided US incomes into 5 equal pieces called “quintiles”.  Now, these income intervals will be our sub-populations under name “strata”.  Proposed formula is one that averages the sum-total of each stratum’s percentage proportion in total population multiplied with its mean income.  It is denoted as …

= S (m_i X W_i) / S W_i

      where

      i: Stratum, and i є {1,5}

      m: Arithmetic mean of all incomes in a stratum

      W: % share of the total number of people who earn the

         corresponding income, and 0 < W < 1

From Table-1 (Appendix) that shows percentage share of each stratum, and Table-2 that shows average income in each stratum between 1967 and 2001; (weighted) average income per household for year 2000 is…

= [(1^st stratum’s mean income X 1^st stratum’s % share in total distribution) + (2nd stratum’s mean income X 2^nd stratum’s % share in total distribution) + (3^rd stratum’s mean income X 3^rd stratum’s % share in total distribution) + (4^th stratum’s mean income X 4^th stratum’s % share in total distribution) + (5^th stratum’s mean income X 5^th stratum’s % share in total distribution)] / Total % shares

= [($10,157 X 0.03) + ($25,361 X 0.08) + ($42,233 X 0.14) + ($65,653 X 0.23) + ($142,269 X 0.49)] / 100%

= ($304 + $2,028 + $5,912 + $15,100 + $69,711) / 1

= $93,055 per household

Total income, then would be this average income per household multiplied by total number of households[9].

= $93,055 per household X 105,500,000 households

= $9,817,302,500,000

Finally, per capita income would be total income divided by total population[10].

= $9,817,302,500,000 / 281,400,000 people

= $34,887 per person.

CONCLUSIONS

Average income generated by the proposed weighted average formula supported the hypothesis of this study.  Average US income computed ($34,887) is smaller than the traditional GDP per capita figure for year 2000, which is $35,100.  Finding is in compliance with the argument that GDP per capita overestimates the real average income.  A crucial analysis now is the historical difference between these two figures (past US average incomes announced and the ones that would have been found if weighted formula were used in those years). Income figures[11] between 1967 and 1992 promises an increasing difference for following years and the future:

Graphically…

As this graphics and pure mathematics logic merge, this assertation concludes with below statement: “Due to the highly unequal distribution of US national income, GDP per capita has been increasingly over-estimating the real average income in America.  A new measure, such as “weighted arithmetic average of GDP” proposed in this paper, is likely to give more accurate figures for the nation’s average income.

APPENDIX

Table-2: Mean Income of each fifth in the income distribution between 1967 and 2001.[13]

Table-3: Given and computed income data between years 1967 and 1992

[1] Source: CIA World Factbook 2003

[2] http://www.lcurve.org/

[3] David Schweickart, “After Capitalism: New Critical Theory”. Rowman & Littlefield (2002).

[5] “IISP report-2003”, 5^th International Conference of the International Society for Life Quality Studies (2003).

[6] “IISP report-2003”, 5^th International Conference of the International Society for Life Quality Studies (2003).

[7] Source: World Bank 

http://www.worldbank.org/poverty/data/2_8wdi2002.pdf

[8] Source: CIA

http://www.cia.gov/cia/publications/factbook/rankorder/2004rank.html

[9] Source: US Census Bureau

 http://www.census.gov/prod/2001pubs/c2kbr01-8.pdf

[10] Source: US Census Bureau

http://www.census.gov/prod/2001pubs/c2kbr01-8.pdf

[11] Computations in Table-3 at Appendix.

[12] Source of data: US Census Bureau.     http://www.census.gov/hhes/income/histinc/ie3.html

[13] Source of data: US Census Bureau.  http://www.census.gov/hhes/income/histinc/h03.html