It's hard to take this seriously when the author mixes up the basic meaning of GDP.
> For example, when economists think about how “the economy” is doing, they have traditionally focused on total income (GDP) or per-capita income (mean GDP) as the most important measure.
No, they don't. GDP isn't income at all, they're completely different metrics.
And in fact the article never mentions "median GDP" like the posted title suggests, it switches to "median income" because median GDP isn't even a thing that you can calculate since it's not a metric that exists for each individual person (therefore no median).
But I think that the article's author suffers a fundamental failing of understanding of why people look at GDP per capita. It's not because they're trying to understand the experience of individuals within the country at all -- it's because they're trying to understand how rich the country is, normalized to its size.
True as far as it goes, but the fair response to this is: who really cares how rich a country is, normalized to its size? In the
If you care about the country as people, then median is better as you point out. But if you care about the country as country, then total is generally more interesting (to understand, say, military power).
China's a big deal despite not being particularly rich per capita. On the other hand, the leaders in per capita GDP, outliers like Macau and Luxembourg, are not usually countries people are interested in.
It assumes a normal-ish distribution. Mode would be a better single-number measure, but it’s the income x number of people curve that carries the important message.
I'm sorry I didn't make myself clearer. I'm not talking about GDP, but income (which has a distribution curve).
OTOH, we can see that curve in a global context, of GDP per capita across countries. It's a boring curve, very power-law ish with slightly odd peak in the 30K-40K region.
For example, you are trying to figure out which country has a good pro-growth policy regime. You don't want to imagine that India, let's say, has done better at economic development than South Korea, despite India's higher GDP.
GDP per capita is a yardstick for labor productivity. I’m generally interested in the cost of doing business in various places. GDP per capita is a good metric.
Well no, the whole point of measuring country differences in labor productivity is that in some countries spending x hours to drill a hole in the ground is vastly more productive than in others.
The point here is that, if we want to use GDP as a proxy for labor productivity, then including wealth from extractive industries isn’t that useful because it tells us much more about the geology of the country then about the labor force.
It's a yardstick for productivity per capita, from both labor and capital. I mean, I guess that's just saying "gross domestic product per capita" in different words.
And, let's not forget, "how rich the country is", should be in astericks. It's not adjusted for PPP and therefore it's more a measure of how strong the country's currency is. A country with GDP per capita of x10 times higher than another's doesn't mean, the average person is x10 times richer or even x1 as rich as the other country. And, even when it is adjusted for PPP, the basket of goods tends to be skewed towards less relevant purchases, rather than the essentials of life, thus making it somewhat innacurate.
I've closely compared my own standard of living with someone else in a developing country. And, my country's GDP per capita was over x20 her country's. And, I was surprised to find how closely our material standard of living was: it was pretty much even: (taking into account, number and type of appliances, living space, eating out frequency, quality of food, access to education, water and electricity).
Well I'd say that GDP is much more like revenue than income. You're certainly correct that they are related, as income is typically some percentage of revenue (even negative) however the actual percentage is different for every entity, so it somewhat meaningless.
> But I think that the article's author suffers a fundamental failing of understanding of why people look at GDP per capita. It's not because they're trying to understand the experience of individuals within the country at all -- it's because they're trying to understand how rich the country is, normalized to its size.
Completely agree. Generally there are better metrics if you want to look at the income distribution for a country, e.g. the Gini index.
There isn't a great analogy for it on the individual or corporate level. GDP is a measure on transactions within a system. A person can't meaningfully transact within themselves.
So do a million other measurements. I'm pretty certain you could find some level of correlation with virtually any measurement and income levels. Thus, this is an irrelevant statement.
> Gross domestic product is an aggregate measure of production equal to the sum of the gross values ... or the sum of primary incomes distributed by resident producer units.
I guess it's more nuanced, but I don't have a PhD in economics like the author.
I think shric was just remarking that English is funny because you can take the exact same words of a sentence and rearrange them a bit so it means the opposite.
Most sentences can be rearranged to mean the converse. This is cross-linguistically pretty common, depending on which atoms you choose to rearrange. Most sentences can be negated with at most a couple of word changes. This is cross-linguistically about as universal and anything in linguistics is, again depending on which atoms you choose as "words".
English isn't funny, humans are funny, building all sorts of redundancy into their languages, but then reversing polarity on a dime.
Income is a measure of retained value after all exchanges have had place.
If I buy a truckload of ice cream for $1000, sell it for $2000, and pay the trucker $200, the GDP is increased by $1000 + $2000 + $200 = $3200, while my income is merely $1000, and the trucker's, $200. Our profits are even lower; mine is $800, and the trucker's is $200 minus fuel costs and road tolls.
Wouldn't there be a difference in cases where due to foreign ownership and capital extraction some money would get counted as GDP in one country but GDI in another?
GDP in itself is a faulty measure. One easy way to increase it to ask everyone to throw a brick through their neighbours window. The spend on window repair will increase the GDP, although there is no value add to the economy, other than the glass fitters of course.
Only if you assume everyone is sitting around doing nothing waiting for window repair work to arrive. More likely is that the effort and resources required to fix the windows will displace other economic activity which would have contributed positively to GDP numbers.
GDP is a flow measure. If you drain a dam the flow will increase for a while before you deplete your stock. There are stock measures of the economy, though they're difficult to estimate because an asset's value is only truly known when it is transacted. (And even then in a highly particular way.) One way to think about our capital markets is as an ongoing value-of-the-economy machine (or at least a subset thereof).
Yes, and that's likeky to have an impact on later GDP values, as all that has not been spent on better investments than repairing windows.
Except for the UK that has terrible windows and a proper upgrade would lead to serious reductions of gaz imports, that are likely to sting this winter.
GDP is atrocious, "velocity of money" is worse, but really any attempt to measure "economic well being of most people" (direct quote from OP) based on dollars moving through the financial system is problematic. Finance-industry circle jerks? Counted - often over and over again. Domestic labor (especially child care) by a stay-at-home spouse? Not counted. Made-up increases in asset value? Counted. Barter or charity? Not counted. When GDP is the basis, mean vs. median hardly even makes a difference. Many countries have worked toward better measures of human prosperity, but again the US is falling behind.
GDP is a blunt force measurement. A country with a high GDP per capita USUALLY (not always) has a higher standard of living that one with a low GDP per capita.
Taxation figures heavily into GDP - most taxes end up being measured as a share GDP (and why "financial circle jerks" and "made up increases in asset value" are counted as they eventually... but not always... end up being taxed). That is why "economic growth" is so good for the public sector. More GDP generally means more money to work with. Output that is out of reach of taxation (i.e. domestic labor, barters, charity) is also USUALLY not counted in GDP.
Finally, the Scientific American article is a really good read. GDP is not a measure of quality of life, and really never has been. It's just a measure of economic output.
From the title, I thought it was going to be a ranking of countries GDP per capita. Never mind :)
Since I implemented a median estimate while at Packeteer, I'm always amused by how many "educated" people don't even know the difference between mean and median.
Anyhow, "median income" and "median net worth" figures are quoted in the press all the time. There is not some conspiracy to suppress the median.
Usually the issue is collecting accurate income data. There are plenty of places where people do not necessarily report accurate pictures of their income, and so comparison is generally harder.
GDP is the sum of the market values of all final goods and services produced. It can be determined equivalently by measuring production (or added value), income, or expenditure. "Final" is the key word: middlemen and other intermediate steps don't affect the GDP.
Then there is GNI, or gross national income, which is basically GDP with some adjustments for international situations. In most cases, GDP and GNI are within a few percent of each other.
If a US resident (citizenship is irrelevant) buys Chinese goods, the money spent does not contribute to US domestic production, income, or expediture. The income they use to pay for the goods is based on domestic production / expenditure.
The velocity of money is the ratio of GNI to money supply. I don't see how it's relevant here.
Can you please not post like this, or like the comment upthread?
If you know more than others, that's great, but in that case, please share some of what you know so the rest of us can learn, and definitely please don't put down those who know less, even if you find them annoying sometimes.
> Income is not GDP. A US citizen buying a Chinese good does not contribute to US GDP.
It does, just not very much.
GDP is value added. So contribution of anything done in an economy is (selling price - sum of buying prices of components and services needed to make the thing). So if you buy stuff abroad for 80 USD and import it for 20 USD and sell it in US for 105 USD - you increased American GDP by 5 USD.
And if a USA retailer imports a container of some sweatshop-made clothing for $5 each and retails them for $50 each, that's $5 in China's GDP and $45 for USA GDP, because technically the import-to-retail markup is added value.
> nominal GDP is the number of dollars spent in the economy. Nothing to do with income
Dollars spent by one are earned by another. This is captured in the twin calculations of GDP: the expenditure approach and the income approach [1]. (NX and NFI being the border adjustment variables.)
TNI is the sum of all "sum of all wages, rent, interest, and profits." Wages are in GDP because the purchase and sale of labour is a transaction. Even in the expenditure approach, G includes government employees' wages.
Did you click the CFI link? The "What is the GDP Formula?" section is less than 200 words.
> dollars spent on imports do not increase local GDP
No. Hence the NX/NFI border adjustment. (In the income approach we count foreigners' domestic incomes and citizens' foreign incomes instead of exports and imports.)
Not really. Building new houses contributes to GDP, as does renovating it, but if you're buying an existing house the only part that contributes to GDP is the realtor/lawyer fees.
Income has nothing to do with GDP. Maybe do some research before littering the comments with uneducated takes.
And real income/GDP is what matters more in the end anyway. E.g. nominal wage increases aren't beneficial to people if their spending power is in decline
I too am interested in the difference between GDP and Income, and my usual sources (e.g. Wikipedia) don't satisfy.
1) Regarding velocity: if the other person quickly spends that same dollar, unless they spend it abroad, wouldn't it end up as someone else's income, domestically? So why should velocity matter?
2) Request you to elaborate your explanation using the example of a real country whose per capita GDP and per capita Income differ substantially.
If people in the US buy things from China, the income they earn does not contribute to US GDP. One example.
These are basic concepts with clear definitions. Why so many argumentative comments trying to dispute a textbook definition? Look it up first and stop spreading false information
You're being downvoted because you're wrong about textbook definitions and then calling other commenters uneducated and spreading false information. GDP and GDI (aggregate domestic production and aggregate domestic income) measure the same thing both conceptually and in an accounting sense. Discrepancies are caused by measurement errors, because they're different concepts. This is literally textbook macro-economics (e.g. Mankiw Ch 3, or a wikipedia summary [1], or the Bank of Englands layman explanation [2], or NBER [3]).
If people in the US buy things made in China, it is accounted for in both the GDP and the GDI in China by exactly the same amount (and likewise for US GDP and GDI if people in China buy things in the US).
> If people in the US buy things from China, the income they earn does not contribute to US GDP.
The income has already contributed to US GDP, by definition. The person has already produced goods or services and someone else has spent money on them, which becomes the income of the person in question. If the US person then buys Chinese goods, their expenditure does not contribute to US GDP.
You are looking at this the wrong way, from an individual perspective. National economy is about production, not consumption. Income is something that has already been earned, not something you choose to spend. Production comes first, and then expenditure gives it value and turns it into income. The way GDP is defined, the sum of domestic production in a time period is always the same as the sum of domestic expenditure and the sum of domestic income.
If someone buys Chinese goods, they contribute to Chinese production, expenditure, and income.
If you receive an income, someone had to spend money (in this case your employer). So it does in fact contribute to GDP, as your employer spent money buying your service/finished work.
So income you received already contributed to GDP, even if it just sits in your bank account. Now if you go onto spend it, it will further increase GDP, as you spending it will yield an income for someone else. If you leave it in the bank account, it will not further increase GDP, as it won't yield anyone an income.
Additionally, when a US citizen buys a Chinese good, it does in fact contribute to GDP - namely China's GDP.
If it's not clear to you by now, GDP is almost the same exact thing as money velocity, just divide it by the money supply and you have the velocity of money.
Define "save for a rainy day". Are you talking about permanently putting it under your matress, leaving it in your checking account, or investing it? It will eventually be spent, and the only thing that is changing is the velocity. Either way, that is not necessarily correct.
GDP has multiple different calculations:
Y = C + I + G + NX
or
GDP = Consumption + Investment + Government Spending + Net Exports
Another calculation is
Y = C + S + T
GDP = Consumption + Saving + Net Taxes
Saving either through investment and cash hoarding both count toward GDP. The percentage of GDP that is Saving is the Gross National Savings or National Savings Rate.
tldr; In isolation, scenario 1, Investment is $10 as inventory and Savings is $10. Both add toward GDP
> Thus income can not be treated as equivalent to GDP.
GDP is defined as total income, calculated in a certain way. There are also other equivalent definitions.
Income is the endpoint of economic activity. Your income contributes to the GDP when you earn it.
Your spending is largely unrelated to your income. In any period of time, you can spend more than you earn or less than that. Spending creates new economic activity. That activity contributes to the GDP, which by definition means creating new income.
But since, once you spent the $10, they are now somebody else’s income, namely for the person/company you bought from. Which means that now, by simply summing everyone’s incomes, you have an accurate count of GDP, without ever considering how individuals spend their money.
I think the main problem of contention is, confusion about personal income v. national income, the latter of which includes institutions and is the one being used for GDP calculations. Money cannot be spent without being received by anyone, and the moment it is, GDP goes up in the country it’s received in.
I think you missed why I am failing to understand this one. The full context there is:
> If I earn a dollar and don't spend it, it results in a lower GDP than if I spent it. Thus income can not be treated as equivalent to GDP.
It's the second part that I don't understand. Those two sentences read to me like: "Spending decisions impact GDP, therefore income is not equivalent to GDP."
I don't think anyone is claiming that spending would fail to increase GDP. But I still don't see how income is disconnected from GDP, since that spending would be someone else's income.
> In principle, these two series measure the same object.
Measurements have errors, particularly with something as complicated as the economy. The series are correlated because they are different measurements of the same underlying quantity. The correlation is not perfect because they measure that quantity differently, and so have different errors.
There being 3 different ways of measuring GDP (income, expendeture, and production) is not controversial. I am struggling to find any source on GDP measuring methodologies that does not acknowledge all 3.
Put another way, what produced value does not end up as income on someone's balance sheet; and an expense on someone else's?
Yes, a nation's economy is not a closed system, so you need to account for imports and exports, and this accounting manifests differently depending on which methedology you use.
GDP is also only one metric used to measure the health of an economy. No serious economist will look at it in isolation. They’ll look at everything from unemployment rate and inflation to budget deficits and even sociological indicators.
The author is confusing actual economists with armchair “economists” on Twitter.
Another HN trend applies here: the most upvoted comment being a middlebrow dismissal based on a wrong argument, which is still popular with HNers despite most of the replies to it politely pointing out why it's wrong...
(GDP is calculable via an "income method", differs from Gross National Income only in excluding net transfers of incomes overseas, and per capita GDP is routinely used as a yardstick for income comparisons)
You can calculate the total GDP and divide it by the total population. You cannot (as far as I know), assign an individualized share of GDP to each person, which when summed would equal the total GDP of the country, and then find the median of that distribution.
Fortunately, the author isn't proposing to do this, instead stating pretty clearly that he proposes to use an estimate of the median income, not the median contribution to GDP. So it's not really relevant, except perhaps as a critique of the HN headline.
(He also doesn't specifically say that median income = median salary, although taking sales tax revenues and profit incomes and depreciation of capital outcomes into account probably doesn't actually make much if any difference to the median income vs median salary)
though FWIW I don't think it would be impossible to estimate some measure of an individual's "income measure contribution to GDP" based on that individuals' earned income plus imputed contribution to company profits plus sales tax paid; I just don't think it'd be that useful
well, for a start, flows and holdings on foreign investments, interest and monetary flows, international trade, consumption of existing stock (depending on your particular definition of income).
No, it's just wrong to say they're completely different metrics. The Venn diagram between the economic activity measured by GNI (national income) and the economic activity measured by GDP is almost a circle.
US GDP in 2021 was $23 trillion. US GNI (national income) im 2021: was $23.3 trillion.
"monetary flows"
You're right, remittances and foreign aid are income but not GDP.
I think the article is onto something. It might benefits from a little rewording here and there but I believe you're arguing with a strawman here. I might suggest this: https://news.ycombinator.com/item?id=31940277
First define the rich. many people think it's someone earning $500k through productive work for 30 years, not someone with no income who just happens to be sitting on $20m of cash
Comparing medium income with GDP may be a moderately interesting measure of income inequality. Pretty moderately, because GDP is not income: see the famous example of a gentleman marrying his maid and decreasing GDP by that.
The author is just as wrong as the people he criticizes.
A much better measure of “the economy” is median income because that is a more accurate reflection of the economic well being of most people.
This is a problem in many fields including engineering.
A median or average has a measurement error (and just because one bar is higher than another it doesn't really mean it is actually higher).
The distribution is what is interesting and important.
Which gives, start publishing numbers as (x1,x2,x3,x4,x5) at 10,30,50,70,90th percentile or whatever. And give up pretending complex system are simple.
You're technically correct--the distribution carries a lot more information than the average or median alone.
The problem is that many of these decisions are being made democratically, or by democratically elected representatives. They need to be able to explain the impact of their decisions to the average voter.
Having a single number that goes up or down makes communication and coordination a lot easier, and I think it's worth discussing what the best single number might be.
Choosing a single number necessarily removes information about the distribution, though, and which information is being discarded is inherently a political decision. The Rawlsian veil of ignorance (at least a naive version) would have us choose p0 as the meaningful number; a utilitarian with certain beliefs about the marginal utility of money would prefer the distribution mean. Even the choice to use a normalize GDP per capita is taking a strong stance on the repugnant conclusion.
If I really have to go with a single number I’d go with a trimmed mean, this comes after a long stint both in academia and trying to generate metrics pipelines in industry.
I’ve used it in many contexts and compared mean median and trimmean in many different distributions (even bimodal ones); trimmed mean works amazingly well and is very robust at representing the data’s central tendency in many contexts. Depending on how much you trim it, it can essentially get close to a mean or a median anyway. So consider it more as a mean or median but with a bit less outliers or a bit more extra context respectively.
Or those "elected represenatives" could help educate the people - to e.g. understand distributions.
But instead they educate them in BS, both in schools, and through their government communication. It's more convenient to emphasize BS in curriculums than proper life/political skills...
I think there's a middle ground between having a simple, understandable message and just spouting nonsense. Granted we have a lot more of the latter these days.
Politicians allready talk about e.g. poor people, workers, middle class and rich. So I don't think a concrete group division is too hard for them. Percentiles might be too much to ask though ...
This is stats 101. Mean and median are measured of central tendency, and the median is generally better for skewed distributions. The 5 number summary (or your 5 percentiles) gives slightly more information, but ultimately you are summarizing a complex distribution with a few simple stats. There are upsides (it’s fast) and downsides (it’s incomplete) to doing so. The real problem is how few people understand stats 101, which makes them easy to mislead.
And have the news outlets talk intelligently about what the numbers mean.
The BBC website does something nice where they often have an "Analysis" section under the news. A sentence or two about how to interpret the number would be useful.
If they think we need analysis to understand what is happening in Ukraine, I think we probably need it for interpreting economics numbers. "We" being ... most people. And it wont always bet stats 101.
Who Is This Medianist?
Jonathan Andreas is an Associate Professor at Bluffton University who received his Ph.D. in Economics from the University of Illinois at Chicago.
Ok, cool, not some armchair tech bro that spent 5 minutes writing up some blog.
Most people use life expectancy at birth to believe that everyone died in their 30s in ancient times, even though these numbers are skewed because of huge infant mortality.
Introducing something like life-expectancy at birth, 5, 10, 20, 40 years would paint a much better picture, but it takes hard work and a belief that your audience is not stupid.
It is even more complicated. eg. to understand state of healthcare using life-expectancy at 20 or 40 may provide wrong info. Because it ignores fact that most of the sick and weak individuals have already died by then.
Yeah, you get more information the more data point you explicitly define. The challenge is reaching a compromise between "easy to explain" and "captures a lot of information". Depending on your audience, you probably want to expose more granular data, but for the laymen I think multiple points all on the same function is they best compromise, since you just need to explain one simple function and 4-5 well picked values for that function.
Distributions give us richer data. But even distribution of incomes (in the article) or spending (GDP) aren't terribly rich. They don't answer why. So we can add dimensions to what we measure or fall back to just so stories. And that's why most people don't trust economists - oversimplified models and just so stories.
Sure. I am not arguing that a median is not a median.
By measurement error I really meant noise and disturbance, as defined in control theory. E.g. it is perfectly possible to have a value "average tax in county" with zero noise as in perfect value to the cent, but year to year variance due to disturbances.
There are many mappings from the English word “average” to maths: arithmetic mean, median, geometric and harmonic means. We could be here all day.
But “median is better than arithmetic mean for knowing if a lot of people are struggling” should be uncontroversial.
We’ve rigged up a system that is a race to the bottom on wealth inequality. People have tried to run systems like this before and it always ends in tears.
The interesting question isn’t how to parse the maths, or even if we have a bad, bad problem. The question is, can we and/or will we do something about it soon enough that Western civilization lasts for climate change or whatever to even matter. Because we’re on pace to get into tennis courts and guillotines before much if any of Manhattan is below the water line.
It so happens that the most useful number is usually the dominant (also known as mode or most common value), which is not visible on the histogram, especially for flat heavy tailed distributions such as income these days.
> the most useful number is usually the dominant (also known as mode or most common value), which is not visible on the histogram
Are you kidding? It's the highest point on the histogram; a histogram places very heavy visual emphasis on the mode while obscuring the mean and median.
Sometimes true, but it is the only value that can be estimated without error in some cases.
Neither median nor average is useful for a central tendency in a weird distribution - and income tends to be one.
Histograms typically bucket data potentially hiding the true value of mode or adding error to it.
It really should be given directly, with a count of occurences.
Mode is not the same as modal class - which is what you would get from a histogram.
> Histograms typically bucket data potentially hiding the true value of mode or adding error to it. It really should be given directly
I can't agree with that. It is true that histograms bucket data. But the width of the buckets is chosen to make sense in context. It just isn't relevant if people making $20,176 slightly outnumber people making $20,359. So we put them in the "$15,000 - $24,999" bucket, which preserves most of the information of interest while omitting the information that is of absolutely no interest.
In other contexts, you see other bucket widths. Chinese university entrance exam scores are published with a bucket width of one point. (Some things may be obscured. When I was looking into this, I found that it was very common for the top scores to be bucketed with each other, presumably for some kind of privacy reasons. Beijing was notable in putting low scores into buckets of width ten even in what was technically labeled an 一分一段表 (literally: "one point -- one section" table). Fujian was notable the other way, for preserving the bucket width of one point even for the highest scores; I have trouble believing it's a coincidence that Fujian is one of the academically highest-performing regions of China.)
> Mode is not the same as modal class - which is what you would get from a histogram.
For almost all purposes, modal class is more interesting, more useful, and more statistically valid than raw mode is.
Individual incomes seem like one of those areas where that isn't the case. The modal individual income is likely 0 for most countries (economically inactive people), but the earnings of the people the economically inactive are dependent on will vary hugely.
> Which gives, start publishing numbers as (x1,x2,x3,x4,x5) at 10,30,50,70,90th percentile or whatever. And give up pretending complex system are simple.
How are then the institutions going to sell us the narrative that all is well and growth is always good, once we realize poor people stay poor and rich people get richer?
There is a great book explaning the philosophy, story and problem of using averages in many fields - "The End of Average" by Todd Rose. [1]
I originally found it from the fascinating article on how using averages in the design of military plane cockpit resulted in many pilot deaths. That's the best introduction I've seen about uselessness of averages with multiple dimensions (with more than 3 dimensions it goes bananas). The article is actually is an excerpt from the book, so you can get the sense of the book level from it. [2]
I read that book, and as I recall it takes a left turn into education policy after chapter one. Part 1 sells you a story of personalization over one size fits all, and the next parts tell you how individuals vary, mostly in ways understood by educators, and emphasizing the authors personal story of getting a PhD as a high school dropout and how to remove barriers to educational attainment.
It's probably a fine book if education policy is your jam, and its fine for scholars to publish monographs on their research but I felt a bit mislead given the title and PR excerpts. For example, maybe an examination of Average Treatment Effect and Individual Treatment Effect in medicine, and followed by an adventure into causal reasoning. Or perhaps a treatment on diminishing returns of personalization.
And perhaps if the "how we succeed" portion investigated productivity instead of defining success as getting a degree from Harvard. As software designers we come to recognize that there's diminishing returns on customization in our software, and develop hierarchies of customization -- commonly adjusted settings are easily found, and others may be dotfile / envvar / registry only. There's probably a happy medium between Apple inspired "it just works" philosophy and Gentoo's 'every software custom built' approach.
I tend to explain this with car analogies. If you’re about to drive through an intersection and the light goes orange, you either want to brake hard, or maintain a steady speed, but taking the average of these actions will not end well.
Three economists are on the football field, practicing field goals. The first kicks the ball was to the left, widely missing the goal. The second one kicks and shanks it to the right, again missing the goal. The third one runs off the field shouting "we got it, we got it!"
(When I was younger) My friends colloquially used “orange” to mean a light that was a mix of amber (yellow) and red (meaning “too late to be entering the intersection legally”).
Maybe it's not of much interest to economists, but I find median wealth quite interesting[0]. The data is a bit surprising, especially for USA. For example, it ranks #26 for median wealth but #2 for mean wealth. It also shows that a random middle class Canadian is wealthier on average than a random middle class American[1]. As a side note, the data is probably heavily influenced by real estate prices.
Side A is the irritating number of amateur economics players (me!) and the desire to engage in this topic, because its both fascinating and important
Side B is the extent of "you're wrong" put downs from either other amateurs, or professionals, which is entirely understandable given that the Side A players make collossal mistakes, but which also masks the divides in economic theory: Some people are pretty partisan in the trench warfare across the battlezones here.
So it's foolish to bring maths to the table, because you will be faceplaned by anyone who prefers more complex maths, or who wants to nitpick your maths, or who wants to attack your school of philosophy because they stand on a different hilltop.
Does it mean "mistakes" are acceptable? No. But people can be a bit kind about it, because in the end, we're all talking about the same thing.
Well.. mostly. Not me. I'm stuck in the labour theory of value. all this money stuff is just nonsense.
I never really got the labor of theory of value. It seems like the idea is that value comes from working hard, regardless of whether you actually accomplished anything? What does the labor theory of value have to say about bullshit jobs? How do you decide what efficiency means if you measure value based on inputs rather than results?
It's a bit of a throwaway/joke line because the Labour theory of value does not inform any aspect of modern economic theory on value as exchange. They're probably orthogonal concepts.
That said I wouldn't discount that there is a very fine thesis in analysis of bullshit jobs. Particularly the extreme amount of organisationally justified middle management which inserts mindless beaurocracy into creative acts. Think not just lawyers, but the oceans and oceans of ink applied by junior law clerks to pad out the billable hours and get to the million dollars in costs, but for virtually everything we do now. You think all those 150,000 google employees are creating value?
As a position of quasi religious faith, I think the property of value latent in made things vests in the labour to make them not the priviledge of the capital invested. The alternatives are very ugly: slavery would be the highest form of profitable enterprise when in fact I believe the evidence is strong the other way: you need to pay workers to get value from them. Why? Because they create the value. Invest in slaves, it's less outcome for more expenditure.
It seems like the ugly alternative is that immoral people can profit by not paying expenses that legally and morally, they should pay. But even if it's ugly, that seems about right? They wouldn't do it otherwise.
The ugly reality (if I can riff on your language) is that exploitation and expropriation has been normalised, since we invented feudalism and before. All that changed over time, is how much we understand about the value being made, as it happens.
Most economics is about the practical: if we adjust interest rates this way, what will people do with their money (value) and how does it alter the current balance? They don't much go to the moral side: should we do this. Some do, and typically its in public utility function because in times past, we funded this as tax and spend to common good, eg the build out of power, water but in times recent we've sold this capital investment off, to get short term money, in a belief somehow it will be "better" because "competition" which (again, in my opinion)_is just bogus.
There is a heap of economics in health. Pretty much all the worlds developed economies see that it's better done as a public good, except the USA, and now we have specialist health economists in the USA who work to maximise profit for medical investment. Everyone else can get to "tax is better" but thats not an acceptable truth in the USA.
Nothing to do with labour theory of value. If you think an iPhone is worth $1300 because Jony Ive said so, through Apple, then the cost of inputs come as a bit of a surprise. Why is the use function of a device like this so far removed from its actual input costs, labour included? "rational" economics goes to "it's not worth $1300" very fast because a $25 phone does most of "it" fine. Of course, thats a kindergarden level lie because even ignoring irrational desire, iPhones are a value beyond the cost of inputs, its a classic n-Squared proposition of being able to do iPhone-y things with your fellow iPhone holders. Yet apple stock is worth astronomically more than otherwise productive endevours in mining, engineering, housing. For what? How does apple increase or decrease US GDP in a meaningful way? How does Apple or Google add or subtract from the overall US economy, from CPI, from wages, and cost of housing. Great questions (if you ask me)
If you don't believe in the labour theory of value, then a good thought experiment is to send all the undocumented mexican labour home and see what happens to the US economy. I've read of 10% minimum drop in production across the board, and a massive rise in CPI to deal with food shortages. Thats a pretty high value on labour right there.
If you think of value as being determined by the prices people are willing to pay and not necessarily the actual prices, it’s not much of a surprise that often we pay less for things than their value to us. Paying the full value for essentials like food and water would be very bad and it’s fortunate that they can be delivered for a lot less. (Though, sometimes prices are too low, given the damage that supply chains do.)
In your thought experiment, if the cost of labor rose due to stricter enforcement of immigration laws, we would see what luxuries people can cut back on and what essentials they won’t give up. But the pandemic showed some of this already. Lots of money is spent on luxuries like eating out and travel. (And, yes, iPhones.)
The disconnect between cost of production and price willing to pay for masks and toilet paper brought some of this into sharp focus for me. We're still playing around the edges of almost survivalist levels of concern for things which are not really that hard to source, but somehow have become scarce. And, simultaneously, the lack of cohesion between (almost warring) states means we realize price alone is a poor basis of planning for strategic supply chains: Having every hospital in the USA depend on two factories in China for masks in a pandemic, was no more or less silly than having every computer in the world depend on TSMC in Taiwan and Thai hard drive factories, or Samsung for screens.
We depend on two or three mainly chinese sources of rare earths because the "price" of rare earth mining didn't justify the investment to exploit the other sources we know exist. It will take years to build around this.
In the end, we directed labour to try and solve some of this. (masks) The only people who can be directed these days are either prisoners, or troops. Nobody much likes depending on either to solve labour problems.
Interesting how you only seem to consider the owners of the means of production here.
What is the actual difference for a worker between slavery and wage work?
You can change all the "decorative" stuff but the core concept is still that a restricted elite owns the means of production and extracts more value from workers than it gives back to them. This is how things have always worked through serfdom, slavery and finally capitalism.
As a worker you can now change workplace and thankfully you don't risk physical punishments if you don't work well, but you must work for some capitalist in order to survive, no choice there but a few exceptions.
You must also follow the rules capitalists impose on you on the workplace or you'll lose your livelihood. No physical punishment if you don't follow the rules, but your family is gonna live on the streets.
It could also be argued that "medianism" is the perfect method to hide income inequality, because median is resilient to outliers, so the billionaires will not affect the median at all.
Every metric can be "gamed" and so can the median.
If half of a country were impoverished and half were billionaires, I'd say they have a pretty serious income inequality issue. A median would randomly conclude that either the typical person is a billionaire or in poverty. Best case scenario, the median would equal the mean and give a result that reflects reality, but it still doesn't convey the issue at all.
Median and mean both compress n data points into one, which is only useful given some assumptions about the distribution. Quintiles (or similar) would work for that scenario, but I'm sure there are more complicated situations where they fall short.
Sure, it's a toy example to tease out the underlying shortcomings of the measurement. Societies where the median does not reflect the typical experience do exist.
But resistance to outliers is the point because people don't take billionaires inflating the per capita GDP as "wow, this country has a problem with inequality", they take it as "wow, this country is getting richer".
With GDP per capita, a few oil billionaires and a very small class of well-paid oil workers make Equatorial Guinea look like a middle income country. Median GDP reveals that the average person there lives on less than $2 a day, because Equatorial Guinea is also the most unequal country in the world. But clearly it's more representative to class Equatorial Guinea as very poor (median income) than moderately wealthy (per capita GDP/mean income) or rich (per capita GDP PPP)
Median income isn't perfect because it can hide things like one country having 5% living in extreme poverty and one country having 25% but with typical real world income distributions, an average not massively skewed upwards by how rich the top 10% or 1% is gives you a much more accurate perspective on which country most people have higher incomes in.
I remember having an aneurysm trying to explain this back when Turnbull was PM here.
He said, in a press conference, that the average wage was $75k, there was a wave of lukewarm takes on social media along the lines of "he's lying because I don't know anyone that earns that much". I got yelled at by friends about how much of a shill I was for saying he was 100% correct, but the median wage was around $45-50k.
I ended up explaining it by saying if Danny Devito wandered in a room of NBA players, the average height would still go down.
We are dealing with millions here. To move the needle for 297 million people, the other 33 million have to do a lot.
Assume the 297 million people all make $100k. That's $29.7 million. Now have the 33 million make an even $1 million. That's $33 million. That's 10% of the population making over 50% of the income. Altogether, that's $30.7 million income across 330 million people. Or about $190k per year.
Yes, that's almost double what the 297 makes, but it's far closer to them than to the 33 million.
> If a government regulation benefits Bill Gates $10 billion and every single American $20
I think he meant to say "costs every single American $20", because otherwise, it's technically an ok deal if it benefits everyone even if it benefits one person a whole bunch. And his following statements seem to indicate we should be subtracting $6bn from $10bn.
And I also have a problem with the hypothetical. It's set up in a way to give only one outcome. He couldn't even make it $40 dollars, because then it's $12bn to $10bn. And he has to work with billions because 300 million people are a fucking lot.
But arguing about whether we should be measuring averages one way or the other is kind of missing the real point. Are the people taking the bulk of the income deserving of that income? For the most part, the people taking the lion's share of the income are only doing so because they have put themselves in the position to determine how the profits of an enterprise are allotted. And it's clear that they are not allotting the profits based on the value the employees are bringing to the company.
They don't necessarily deserve it. If I control how an asset is allocated due to some arbitrary event, it does not follow that I deserve whatever I've allotted to myself. It can be "legal and above board" and still be unjust. Because the laws and boards can be constructed to allow what I'm doing.
Also, define bribe. Define corruption. If my best friend is a lawmaker and I spend every weekend with them and I manage to persuade them due partly to the amount of time I get to push my point view, is that corruption? Is it corruption if I become friends with him solely because he's a lawmaker? If I didn't know before I became friends. If we were friends before? Why are any of these different, if any?
I like how you say it's not clear that they aren't allocating profits based on the value, but somehow it's very clear they deserve it, what corruption is, and what bribery is.
No. The only thing that is not clear is that maybe they don't deserve the sheer amount of wealth they've extracted from everyone else.
> If I control how an asset is allocated due to some arbitrary event, it does not follow that I deserve whatever I've allotted to myself.
i take issue with the idea that an "arbitrary event" is the reason why such an allotment is considered undeserved. This isn't any different than a lottery winner being considered undeserving of their winnings.
If one traces far back enough, you could argue that the current western country's wealth is undeserved, because way far back, an "arbitrary event" (aka, conquest and dominion) gave them the kick-start that nobody else got.
I would instead argue that no event is arbitrary - the lottery winner chose to make their bet, the lobbyist deliberately attempt to influence the law to their favour, etc. If such actions are not considered illegal by the laws of society at the time, then it is fair game. The good thing about western society is that laws aren't set by a dictator, but instead can be changed via some sort of consensus procedure.
The only argument i hear that such wealth is undeserved is that "i couldn't have done the same, therefore it must be undeserved".
I mean, technically, lottery jackpot winners aren't deserving of the jackpot, no.
And I see you're getting it. There's way more luck involved than most people give credit for.
It's undeserved not because no one else could have done the same, is that because a lot of people could have done the same given the same circumstances.
Take your lottery example. If you buy one ticket, you have one chance. If you buy 50,000 tickets, you have 50,000 tickets. But you first need $100,000 to do that.
And the laws aren't being changed by some sort of consensus... of people who already have the most. And they change those laws to ensure they get to keep the most and anyone else who would like to get ahead have a hard time doing that. The large income inequality makes socioeconomic mobility more difficult.
>Assume the 297 million people all make $100k. That's $29.7 million. Now have the 33 million make an even $1 million. That's $33 million. That's 10% of the population making over 50% of the income. Altogether, that's $30.7 million income across 330 million people. Or about $190k per year.
I understand GDP as a tool that was originally built to quantify how much a nation is producing, and in result focus the economy to increase that number.
But like many tools, at some point they expire and we need to move on. imo GDP is past that point already because it encourages production beyond anything else, when we should instead focus on consuming less and producing enough, and the right things so that we can be satisfied in our lives, not just burning gas for the sake of it.
To illustrate the obsurdity of it, the production of oil contributes to GDP, the burning of that oil contributes to GDP and then using carbon capture to capture the gasses released by the oil also adds to GDP. What's the sense in encouraging all that without qualifying that it should be enhancing people's lives?
Happiness is a complex, multi-faceted thing to measure and is not a good direct target. You want more specific, granular goals that are important for those. Things like health, wealth, social connections, leisure, education, etc.
And GDP is a decent metric to track the economic aspect. It's not perfect, but nothing is, and the answer is to find and improve the metric instead of "throw away and ignore". And, the data clearly shows https://ourworldindata.org/grapher/gdp-vs-happiness GDP has a very strong correlation with happiness.
Economic policies should take GDP into account, and suggesting NOT to do so is simply stupid. The policy makers should also look at and incorporate climate issues, and many others.
> Fetishizing the GDP will only accelerate the climate problem.
I'm afraid this is not a useful framing, and mostly a straw man argument that makes the discussion unproductive. Most countries and their politicians are not pursuing "GDP at all cost" type of policies.
>Don't even look at GDP, look at happiness of people instead.
"What is love? What is creation? What is longing? What is a star?" -- so asks the Last Man, and blinks.
The earth has become small, and on it hops the Last Man, who makes everything small. His species is ineradicable as the flea; the Last Man lives longest.
"We have discovered happiness" -- say the Last Men, and they blink.
Happiness isn’t a great metric because people can be happy dying from tetanus after stepping on a nail because they don’t know it possible to not die from that.
Or even better, forget about GDP and use national income instead since it is better for measuring sustainable growth. For example if a country takes 100 billion euros of oil from its underground reserves, it will have a 100 billion GDP but 0 euro national income. We need to improve the economic indicators that use or at least be aware of the implications of our choice since they impact economic strategy
But economic statistics are not made with the goal of "finding out how average Joe is doing", but to answer the question "what't is in it for me? where i should invest? and in which instruments?". It's audience is private investors and fund managers. Not even politicians - they live in post-truth world, they need to know how to better manipulate people into believing what they need them to believe, what's actually going on is of secondary importance.
And for this, we need means, not medians. Sure vast majority of profits is made from consumption of the top 10%. Rest are just taken for a ride (not even exploited properly, and this is a big departure from things were 100 years ago when the rest were equally desperate, but at least necessary).
Median income of the total population or "the labor force" or of the "potential working age population"?
With barely over 50% of the population working in most countries [1] median based on population will be skewed too.
Then we have to agree on working age population.
Then we have to measure unmeasured income (ex. providing child care, teaching at a home school, preparing meals, cleaning, maintaining your own car, repairing or upgrading your residence, etc.) oe you lose all of this in the data.
The real challenge from my perspective is trying to boil down something as complex as a population into a single stat.
>This simply applies democratic norms to economic questions. Democracy prioritizes the preferences of the median voter, and medianism extends this idea to economics.
Democratic norms are what produced this recession.
“When the people find they can vote themselves money, that will herald the end of the republic.” -Benjamin Franklin
This recession the USA just entered is the same recession as the one which started in 2020. The government can print significant unexpected money to temporarily boost GDP.
Except only about 2 trillion made it to gdp. The US government's leaky bucket can be measured at about 28%. 72% of US tax money is going into the firepit.
It worked, they temporarily boosted the GDP numbers to pretend like there wasnt a recession. When they stopped printing money, gdp dropped because the recession never left.
You know what I really wish to see produced for all countries?
I've actually had this running theory that all civil servants should be paid a fixed percentile of income. This would give them all an interest in ensuring that income/GDP goes up and that it goes up for those below them. (in percentile)
Unless it was a very high percentile, that would incentivize competent people to flee to the private sector, or it would encourage competent but corrupt people to join the government.
A better compensation mechanism would be to measure changes in outcomes such as wealth, crime, and health, then have bonuses for positive changes in those metrics. For example: If each member of congress got an extra $10 million if homicide rates went back to 2019 levels, I guarantee we'd fix the crime wave within a year. It'd only cost $5.4 billion, or $15 per US citizen. And it'd avoid approximately 5,000 murders per year.
One can imagine similar incentive strategies for health and wealth. Unfortunately, most people view governance as a sacred value, and are unwilling to introduce money as an incentive. So we're stuck with things like congress giving the FDA more power to restrict nicotine vaping products while 400,000 Americans die every year from smoking tobacco.
You could pay them, say, 3x the 20th percentile salary.
(I'm not sure it's _civil servants_ this sort of thing should be considered for. They don't make the decisions and there's little a civil servant can do that will change what the 20th-percentile person gets paid. It would make more sense for politicians.)
I suspect that if each member of Congress got a $10M reward for bringing homicide rates to 2019 levels, the result would be legislation that redefines homicide or changes how it's reported or something of the kind, rather than anything that actually involves fewer people killing one another.
This is a solved problem, as it comes up in situations such as CEO bonuses, sports betting, and prediction markets. You either specify the definition in the bonus (eg: according to the FBI's UCR data definitions as of 2019) or you delegate an arbitrator (eg: the GAO). If this weren't solvable then the world would look very different. CEOs and acquired companies would never have compensation tied to target metrics. Sports betting would be rife with lawsuits and fraud. And economic metrics would be as accurate as those claimed by the Soviet Union.
> A better compensation mechanism would be to measure changes in outcomes such as wealth
I agree that GNH (gross national happiness) etc matter too, but I think that tying their income to lower percentiles would greatly increase the correlation between GDP growth GNH... the only reason it's broken in the US is because GDP goes to the upper ranks, in a rent seeking manner, from the lower ranks.
> Unless it was a very high percentile, that would incentivize competent people to flee to the private sector, or it would encourage competent but corrupt people to join the government.
I actually think there needs to be a return to the "service" mentality to civil service. IMO government shouldn't be a job you do because of the compensation, stability, union etc. It should be a service to your country. But that would require so many Americans to stop hating their nation as they seem to now.
Is there any evidence that having higher income makes people less greedy or self-serving? Most legislators are already wealthy or at least well-off before they get elected.
Looking at a measure at centrality implies you understand the distribution is represents. I find it good practice to either observe the distribution before any derived measure, and if that's difficult for some reason, get a few measures of centrality, mean and median if you can. The further they are apart, the unlikelier it is you're dealing with a normal distributed quantity, an assumption people make far too often to my taste.
Reading the comments here, it seems to me that the problem is more that we don't illustrate these types of concerns correctly. I mean we don't bother with pictures, animations and simple stories, but instead argue over too simple or too complex statistics. I love math but it's not an effective way to communicate important ideas. All the discussions here are discussions about statistics minutiae and arguing about that will never change anything.
There is no average that does not hide information.
To point out the obvious, the perfect information is the underlying data. If there are 10 people in the economy then just provide those 10 data points. If there is a million people, then provide all the statistics - the averages, percentiles, etc - as well as all the data.
An article can report the averages as well as point to the underlying data set.
The idea of pronouncing one average better than all the others is moronic.
How is that even calculated? The GDP is the sum of all goods and services produced.
I guess when Amazon sells an item the difference between the wholesale price they pay and the end price they charge including shipping is a service, retail.
But who has produced this service? The underpaid warehouse worker and delivery driver or Jeff Bezos?
I'd say income and debt are more interesting than GDP. The Average Joe in my country saves a bit of money, burns it quickly, then struggles for a while because he needs to borrow money, and he doesn't make enough to pay it quickly. Many people are blissfully unaware of how limited their runway is, particularly when inflation hits hard.
GDP is a very weird metric. For example, GDP includes money spent for renting an apartment, although no goods are produced in this case.
So for example, if there is a village with two people, and they rent each other's home then this village has non-zero GDP, although no goods or services are produced.
Inflated education, medical or insurance prices also inflate GDP. So a teacher in one country could produce 2x more GDP per unit time as the same teacher in another country.
Another wrong thing to do is to adjust income or GDP for food prices (purchasing power parity). This is absolutely wrong. The idea of PPP is that if in your country potato is two times cheaper then it makes you two times richer. This is wrong because a MacBook or a car are not two times cheaper. Actually they are more expensive than in the West. And if you go abroad then there will be no cheap potato either. So PPP is a total lie. It can be applied only to poorest people who spend all of their income on potato.
So if you want to adjust the income for PPP, at least use MacBook or car prices instead of potato.
I think that a fair metric would be a percentile distribution of how much money a person has left after renting the cheapest apartment and buying the cheapest food.
PPP isn't based on "a potato", it's typically calculated based on a representative price index covering all the different types of goods people buy.
Theoretically it would be possible to create a price index based on Macbooks or cars, but since most people in the world buy more food than Macbooks or cars this would obviously be misleading. When people living in developed countries move to a developing country they don't think "wow, the cost of living is high, the Macbook has a luxury tax on it" they think "wow, most things are incredibly cheap here". It is impossible to come up with a perfect PPP because people spend their money on different things, but it's a lot better than thinking a country which has inflated prices for everything is much more materially wealthy than one which has low prices for everything.
And the whole point of PPP adjustments is to get rid of the scenario where the same education contributes much more to GDP in a nation where wages and school buildings are expensive than one where they're not.
Well, yea, but you can't calcuate GDP per capita. You have to instead look at median household income. If you wanna sell fancy electronics in mid market economies it really helps to know the difference.
Seriously how do you measure the median of GDP per capita? Do you measure the GDP output of each person of the population and take the middle person's GDP outpt?
> For example, when economists think about how “the economy” is doing, they have traditionally focused on total income (GDP) or per-capita income (mean GDP) as the most important measure.
No, they don't. GDP isn't income at all, they're completely different metrics.
And in fact the article never mentions "median GDP" like the posted title suggests, it switches to "median income" because median GDP isn't even a thing that you can calculate since it's not a metric that exists for each individual person (therefore no median).
How did this get to the front page?