M2. Why is MELI Better Than GDP?

MELI (Median Expected Lifetime Income) should replace GDP per capita as the main measure of economic welfare.  Simon Kuznets won the third ever Nobel Prize in Economics for his many years of work developing measures that ultimately evolved into Gross Domestic Product (GDP) because GDP really was revolutionary.*  Kuznets did not originally intend for GDP to be used as a measure of welfare.  Kuznets wrote that “the welfare of a nation can scarcely be inferred from a measure of national income.”  But with the rise of mmutilitarianism, GDP was divided by population to get mean GDP and that soon became the primary measure of national welfare.   GDP is a measure of production (which is approximately equivalent to income), and there are many problems with using it as a measure of welfare.  Thousands of people have tried to create alternative measures to replace GDP, but no alternative measure has come anywhere close to replacing it.

One problem with all the alternatives is that they are more complex than GDP and that makes them harder to interpret.  Sometimes the additional complexity introduces more measurement errors and other times the complexity just makes the measure harder to understand what is actually being measured.  By far the most successful alternative has been the Human development index (HDI) which suffers from both problems and it is nowhere near as common as GDP as you can see from comparing how often the terms are used in the written English language.

One Measure Gross Domestic Product

I think the above comparison greatly overstates the importance of median income and the HDI relative to GDP because every time new GDP data comes out, it immediately influences important government policy about monetary and fiscal affairs whereas I have never heard of anyone changing any policy in response to the other two measures.  Furthermore, people are so familiar with GDP that we rarely use its full, formal name.  We prefer to use its more intimate nickname, ‘GDP’, which is used more frequently than everything else put together.  This graph is a better representation of how much more power GDP has compared with the other measures.

One Measure To Rule Them All

Because the HDI is the second most important alternative measure of welfare, I’ll compare the HDI to Median Expected Lifetime Income (MELI) later but note that it is already a less common concept than median income according to this data.  Furthermore, GDP is the most important component of the HDI, so the HDI doesn’t completely escape from the cult of GDP.

Medianism proposes to replace mean GDP with MELI.  MELI still has some of GDP’s problems such as that it does not measure changes in wealth stocks, but it has all of GDP’s useful properties as a measure of welfare and it repairs several of GDP’s problems.  The Complex Systems Group at Bluffton University is working on developing a specific methodology for measuring MELI, but the simplest way to calculate it would be to multiply the median individual income in a given year times the median expected lifespan for that year.  That would generate an estimate of lifetime income for the median individual within a group.

Advantages of MELI over GDP:

  1. MELI data is cheaper and simpler to gather than GDP because whereas GDP requires measuring ALL production everywhere, MELI only requires two things which most nations already do: 1. a registry of births and deaths and 2. a survey of incomes focusing on households near the median.  The first is zero additional cost for most nations which are already tracking births and deaths.  The second is zero cost for all developed nations which already have detailed tracking of household incomes near the median because this is the foundation of income taxes.  Income tax records miss some income such as implicit rents, but they are incredibly comprehensive and could be complemented with additional surveys.  Even for nations without income taxes, surveying households is much cheaper than measuring GDP.  Unlike GDP, there is no need to try to measure the incomes of all the richest and poorest households because only the households near the median need to be measured.  The richest and the poorest are the hardest to estimate because the rich tend to hide their income to escape taxation and the poor are harder to find and are less likely to accurately track their incomes because they have less pressure to track their own income for income tax recording purposes and do more informal work.
  2. MELI is  simpler to explain than GDP because GDP is more complex.  It has only been adopted despite its complexity because of its 80-year history and official status around the world.
  3. MELI avoids counting many parts of GDP that have little to do with wellbeing.  For example, investment and net exports are poorly correlated with the flow of experienced wellbeing compared with household spending.  Separating out the household consumption component of GDP gives a better indicator of economic wellbeing than GDP. For another example, GDP includes the implied value of financial services which have nothing to do with wellbeing.  It is really more like a measure of the risk that banks are taking on.  During the run up to the financial crisis, investment banks produced ever greater contributions to our official GDP measures even though there is no evidence that this benefitted the average household and it turned out that they were creating the greatest financial crisis since the Great Depression.  MELI avoids this problem completely.
  4. Income is approximately lognormally distributed within every nation which means that it is skewed to the right as illustrated in the graph below.  The red diamond marks where the median is.lognormal distribution of incomeFor skewed data like this, the median is a better measure of central tendency (a better ‘average’) than the mean and this makes intuitive sense because the mean is skewed by outliers.  In the example above of the median income of the drunk illiterates if Bill Gates walks into the bar, what would happen to the average wellbeing of the people?  Mean income skyrockets and gives the impression that everyone is a billionaire, but median income only increases slightly.  Median income is a better measure of the average economic wellbeing of the group.
    • Because income is approximately lognormally distributed, statisticians routinely use the log of income for statistical analysis, because when they do that, the income data transforms from the skewed distribution shown above into the standard normal bell curve.  When they use the mean of log income that happens to be the median income!  So statisticians are already using an approximate estimation of median income when they routinely use this statistical trick with income data.  When they use the mean of log income, they are implicitly recognizing that the median is a better measure than the mean of linear income.
  5. Median income deals with inequality better than mean GDP.  The nations of the world with the highest inequality won’t want to publish their MELI.  For example, Monaco and Qatar are ranked among the highest GDP per capita in the world even though most people in these countries are not particularly rich, but they probably would not fare very well using MELI.  The failure to adjust for inequality is one of the main criticisms of GDP as a measure of economic wellbeing, and MELI is the simplest way to fix it.
  6. MELI incorporates both of the most important and most measurable economic statistics into a single measure that makes intuitive sense.  Mean GDP is just a year-long snapshot of economic production, but there is nothing particularly special about estimating income during only a year or a quarter, or a month. People care much more about their lifetime income more than their income in any particular year.  Although MELI cannot really predict lifetime income because it is agnostic about future structural changes in the economy, it would predict the lifetime income of the average person born in the measured time period assuming that economic structures remained constant throughout their future expected lifespan.  Adding a measure of lifespan to our income measure is important because at the margin, the average American cares more about lifespan than about money. We know that because of the revealed preferences in the value-of-life literature. Americans are willing to tradeoff many hours of work to be able to spend their wages on safety and healthcare and safety that only gains a few hours of additional lifespan.
  7. Amartya Sen has said that GDP fulfills two completely different purposes.  First is a measure of productivity and the second is as a measure of welfare.  If we adopt MELI for welfare, then the government could focus on making GDP into a better measure of productivity.  In particular, the government should focus on Net Domestic Product (NDP) which would be a feasible improvement in measuring productivity.  It would subtract depletion of natural resources and depreciation of capital.  NDP is a worse snapshot of wellbeing than GDP, but it is a much better snapshot of current productivity.  Environmental concerns and important and although MELI cannot address them, if MELI took the welfare purpose away from the armies of statisticians who work on annual GDP measurement, that could free them to make GDP a better measure of net production by incorporating more environmental depletion and degradation into that statistic.  This is roughly what Nordhaus and Kokkelenberg’s 1999 book recommends. One roadblock that prevents achieving these goals is that they would weaken GDP’s usefulness as a snapshot of current wellbeing in a given year because current consumption gives a sense of current wellbeing even if it is spending down resources that may lower future wellbeing.  A lot of the environmental adjustments to GDP would show that we aren’t really producing as much as GDP is counting, but GDP does give a more accurate picture of current consumption than the NDP would give.  We should replace GDP with two different statistics that each focus on these two different functions.  MELI is a better snapshot of economic wellbeing and NDP is a better measure of production.
  8. Economists think that they eliminated utilitarianism during the “ordinal revolution” through the works of Lionel Robbins, John Hicks, and others.  What really happened was that they merely banished dimmeu utilitarianism and adopted mmutilitarianism in its place.   But the ordinalists did have a valid point in claiming that money merely gives an ordinal measure of utility.   Unfortunately, they did not follow their own logic to its natural conclusion which would have led to MELI.  If money is merely an ordinal measure of utility, then it is impossible to add up different people’s money and use the mean as a measure of their average utility because that would require cardinal (interval scale) data.  For ordinal data, the best measure of central tendency is the median as every basic statistics textbook explains.
  9. Libertarians should prefer MELI because ever since the beginning of GDP statistics, they have complained the value of government production is overstated and some have gone so far as to argue that no government expenditures should be included in measures of production and wellbeing at all.   The micro-data that is used to calculate median income excludes government production in most nations by default because all household data series collect information about market incomes and only some even include information about taxes and transfers from the government.  In the future, a second version of MELI could develop to satisfy people who will want to account for how much government is contributing to the average citizens relative to the taxes they pay.  To account for this, there can be one version of MELI that estimates lifetime income before taxes and transfers and another that recognizes what the government does to real household consumption.
  10. MELI is more multi-dimensional than GDP because it also incorporates lifespan which is very well measured and extremely important for wellbeing.  Many developing nations looked like basket cases in the 1960s and 1970s because they were judged by stagnant mean GDP.  However, their life expectancy was dramatically rising which was inflating their population and it was difficult for the production of goods and services to keep up with the production of babies.  GDP greatly understates the increase in wellbeing that greater longevity brought.  In contrast, MELI much more accurately reflects the growth in wellbeing that was brought about by longer lifespans.


*Kuznets developed the methodological foundations of GDP, but at the time he actually used a slightly different measure, Gross National Product (GNP), which divides income between nations a bit differently from GDP, but which is the same on average across all nations.  GNP is a measure of the income of a nation’s people and GDP measures the production within a nation’s borders.  The difference between the two is when people in a nation get net income from foreign possessions.  This is fairly trivial for the US which is why there is little importance whether Americans choose to use GDP or GNP (GNI).  Today GNP has largely been replaced by GNI (Gross National Income) which is a better descriptor.

Back to Medianism FAQ