The Growth Illusion

“For over seventy years, the advanced societies of the world have preened themselves in the mirror and mostly admired what they saw: growth.” So starts The Growth Delusion by David Pilling. Its subject is Gross Domestic Product, or GDP, and the importance our societies have put on growing GDP. Indeed, news outlets never miss a chance to report on the latest quarterly growth figures, while politicians often run on their stellar record of ‘economic growth’, or blast their opponents for disappointing figures. In some ways, “GDP has become a proxy for a country’s well-being,” Pilling says in his book The Growth Delusion.

As defined by Investopedia, GDP is “the monetary value of all the finished goods and services produced within a country’s borders in a specific time period.” By specifying that it measures ‘finished’ products, intermediary products that went in the making of the final good are excluded, because not doing so would lead to double counting. For example, the baker bought flour to make the bread that he sold. Baked within the price of the bread is the cost of the flour. Therefore, counting both the bread and the flour would lead to counting the flour twice. GDP is also limited to a region, typically a country. Finally, GDP is measured over a period, typically a quarter or a year, because it tracks a flow of transactions rather than a stock of assets (i.e. wealth). Going back to our bakery, GDP would measure the flow of transactions, or revenues. The stock of the bakery’s assets would be the invested capital (e.g. ovens) and cash in the bank. In summary, GDP measures the monetary value of all the stuff that is produced in a region in a period. That is why people often equate GDP with ‘the economy’.

Yet as shown by the David Pilling book, The Growth Delusion, GDP is flawed. And if GDP is to be widely used, we must understand its limitations. This essay will attempt to do just that. It will be divided into two parts. Firstly, I will review Pilling’s book, which sets solid foundations for better understanding GDP and its shortcomings. This section will cover the common criticisms of GDP and try to make sense of them. Secondly, I will discuss my own thinking on the subject, a lot of which occurred when reading The Growth Delusion. Although I did not expect to come to such a drastic conclusion, I have come to believe that GDP growth is not a delusion, but rather an illusion. It does not exist. But I am getting ahead of myself. Before going through the looking glass, let’s see what the experts have to say on the subject.

Part 1: The Growth Delusion

Imaginary boundaries

Most people alive today cannot remember a time before GDP. It might seem like it has always been with us. But it has not. Although Pilling reminds us that people have for a long time tried to measure the size of the economy, often to estimate how much funds the monarch could raise to wage war, the first comprehensive account of ‘national income’ was produced in 1934 by Simon Kuznets and his team at the US National Bureau of Economic Research. President Roosevelt, who had hitherto relied on anecdotal evidence such as stock prices or freight-car loadings to assess the then-raging Great Depression, would now have an estimate of the value of all goods and services produced over a period of time in America.

Quickly, debates on the methodology of computing these national accounts raged. What should be included? Kuznets wanted government expenditure out along with illegal and harmful activities. Others, like John Maynard Keynes in England, wanted government spending in. Keynes eventually won the day, probably because governments like to think of themselves as contributing to the economy, but the broader debate did not end there. The discussion of what would come to be known as the ‘production boundary’, which decides what goes in GDP and what does not, has carried on ever since.

In 2014, the United Kingdom for the first time included prostitution and drugs into its GDP figure, boosting GDP by £10 billion, and causing the nation to become the fifth largest economy in the world, overtaking France. But was the UK suddenly that much richer? No. Nothing had changed for Britons. The decision to include the labor of prostitutes alongside that of accountants, grocers, and engineers led to a redefinition of the production boundary, but it did not materially improve anyone’s living standards, except perhaps that of the statistician tasked with surveying that new industry.

The trend historically has been to include more of what economists call ‘marketable goods’, meaning that money is exchanged for a good or service. Drugs and prostitution, no matter how unpalatable they are, conform to that standard, but not all activities do. Many things produced in an economy are not bought and sold on the marketplace. As Pilling explains:

“If a Japanese housewife cooks her ageing father-in-law meals, helps him in and out of bed, helps him use the toilet and washes his clothes and sheets, none of her efforts count towards the economy. If, however, she works in a care home looking after someone else’s father-in-law — and earning a wage while she’s at it — then the exact same activities contribute to national income.”

Quite evidently, the current consensus to only include marketable goods and services can lead to curious outcomes. Paul Samuelson, an economist and Nobel Laureate, once quipped that a man marrying his maid would decrease GDP, as her work would now move on the other side of the production boundary. The same amount is produced but suddenly the ‘economy’ has shrunk. In fact, it has not. The economy is simply not capturing all production.

The danger, as Pilling notes, is that, “if we do not measure something, it is undervalued.” For example, including cooking, cleaning, washing, and driving into GDP would, by some estimates, grow the US economy by 26 per cent. If these numbers were reflected in national accounts, perhaps policymakers along with the public would more readily recognize the value of stay-at-home spouses. The sad reality, Pilling points out, is that, “policymakers and regulators are biased towards what they can see and what they can count.” Therefore, GDP is the arbitrary result of what is included in the production boundary and will undoubtedly end up causing prejudice.

Stocks versus Flows

Looking at GDP, China appears wealthier by the day as sleepy countryside villages are turned into thriving cities. Yet breakneck economic development has drawbacks. As Pilling reminds us, “two-fifths of China’s river water is undrinkable and one-sixth is so polluted it is unfit for any use,” and “according to one estimate, soil is eroding in China thirty to forty times faster than it can be replaced naturally.”

The fact that GDP only measures production over a period (i.e. a flow) and fails to account for the resources or assets that were consumed in the process (i.e. a stock) can lead us astray. Pilling explains:

“Imagine for a moment two people, Bill and Ben. Bill is a banker and earns $200,000 a year at Goldman Sachs. OK, he’s miserably paid by banking standards, but bear with me. Ben is a gardener and earns $20,000 pruning roses and trimming hedges. Who is better off? If you measure the income each receives, then Bill is clearly richer, in fact precisely ten times richer. This measurement is the equivalent of GDP; it tells you about the ‘flow’ of income each receives in a year. But, just like GDP, these numbers don’t reveal much about the true wealth of Bill and Ben.

To discover more, you’d need to know about their stock of assets. Did I forget to mention that Ben the gardener recently inherited a huge country estate in Long Island worth $100 million? In truth, he works in his own vast garden as a bit of a hobby on Tuesday afternoons and pays himself a token wage. But he plans to sell off the estate next year, move into somewhere more modest in Manhattan and live off the interest from investing the $95 million or so he’ll have left over.

Poor Bill, meanwhile, is up to his neck in debt. He has to fork out half his salary each month on his mortgage, which has another ten years to run. He has car payments on his (scratched-up) Porsche and a troubling bank overdraft that he’s acquired to maintain his highfalutin lifestyle. Unfortunately, he’s also pushing fifty (Ben is nineteen by the way) and the bank is going to have to let him go.

Now who looks better off, Bill or Ben?”

The conclusion seems clear: ignoring the assets, or stocks, required to generate GDP is foolish. A simple way of picturing this is with a bathtub. If being wealthy means having a bathtub full of water, you cannot simply focus on how fast water flows out of your faucet (i.e. GDP). You must also think about how fast water is coming out through the drain (i.e. the depletion of assets). Businesses understand this. Depreciation is the accounting term for the cost of ‘wear and tear’ on equipment used to manufacture a product. If the depreciation cost of making one widget is higher than the revenues that it can generate, it is not worth making. Yet in national accounting, it seems common practice to ignore depreciation.

Public infrastructure, such as roads, require constant investment to counterbalance depreciation or wear and tear. A country with world-class infrastructure could afford to underinvest in its infrastructure for a few years without facing much problems, but it would not be sustainable. If, say, 10 percent of a country’s output goes to maintaining its existing infrastructure, 90 percent remains. If however 20 percent of output is required just to stand still, that is significantly worse. Yet GDP does not allow us to differentiate between these two scenarios. This limitation should not come to policymakers as a surprise. The hint was in the title. ‘Gross’ in ‘Gross Domestic Product’ means before wear and tear. A more honest figure would therefore be ‘Net Domestic Product’, which includes the cost of depreciation of capital goods, which includes housing, vehicles, and machinery.

But NDP does nothing to solve the problem of depleting natural assets. Depreciation is the wear and tear on stuff that humans make such as cars, roads, or factories, but it does not account for depleting topsoil or air quality. So, for example, a country with large oil reserves might look like it’s prospering from a GDP standpoint, while in fact it is heading for economic ruin in less than a decade when it runs out of oil. Or a country could increase agricultural output every year for a decade, only to realize that nothing will grow on its depleted soils afterward. GDP does not tell you how sustainable that flow is.

The conclusion, in all these cases, is to remember that assets are required to generate income — whether that be a factory, oil reserve, or topsoil. GDP does not tell us the cost we incur by depleting assets when they are used to generate income and therefore what can sometimes look like a prospering economy might in fact be one heading for ruin.

Digital Free Lunch

Consumers, empowered by the digital revolution, have increasingly been doing themselves what used to be done by professionals. We scan our own groceries at the supermarket and book our own plane tickets online. Whereas once people paid for encyclopedias written by professionals, now Wikipedia is free and articles are written by volunteers. As a result, parts of our economy which used to be included in ‘marketable transactions’ are moving to the other side of the production boundary and thus out of GDP. As more transactions shift to the order side of the boundary, GDP captures less economic activity and therefore shrinks, despite technological innovation making our lives significantly better.

Secondly, GDP struggles with quality. GDP growth measures the increase in the quantity of goods and services produced, but it fails to account for improvements in quality, such as that arising from technological innovation. If the number of computers made increases from 100 to 105, GDP will record a 5 percent increase. But what if technological innovation had made the new computers twice as fast as the old ones? Standard measures will not pick that up and adjustments are therefore needed. In 1996, the Bosking Commission, which had been appointed the year before by the US Senate to investigate bias in inflation estimates, concluded that because of “rapid advances in equipment like computers and phones, the US had been overstating inflation by 1.3 percentage points a year before 1996,” Pilling writes. “That meant it had been understating growth by the same amount.”

Unfortunately, The Growth Delusion does not go far enough on this topic. Perhaps gradual improvements in quality such as computing power becoming ever cheaper can be accounted for with adjustments because it is essentially the same product becoming much cheaper, but a lot of innovation does more than that; it creates something entirely new. How would we account for the discovery of a new drug? Or the jump from flip phones to smartphones? This form of innovation causes a rupture. It is not a cheaper version of the same product. It is a fundamentally different product. GDP cannot account for that form of technological innovation. This argument will be further discussed in the second part of this essay.

Two Tickets to Paradise

People often use GDP to compare the relative wealth of countries. We can for example say that Americans are on average thirty times richer than Indians. By this, we mean that Americans could buy thirty times as much stuff as Indians or that for every $1 earned by an Indian, an American earns $30. But give an Indian $1 and an American $30 and you will quickly realize that $1 goes a long way in India because of cheaper labour. Conversely, no matter how much the Indian spends, she will not be able to buy cheap Texas land and import it into India to live on. In simple terms, some goods and services cannot be traded. In his book, Pilling, who lived in Japan, talks about the punctuality of Japanese trains. No matter how much he is willing to pay, he cannot get a Japanese train from London to Manchester. Some goods, and especially services, are not traded between countries. As Pilling explains, “[i]n the context of non-tradable services like a train between Tokyo and Osaka, the price test breaks down.”

There are large parts of our lives which fit in the ‘cannot trade internationally’ bucket. Infrastructure, such as roads, ports, electricity, internet, and trains, are all local. Services are too. French waiters are rude, no matter how much you tip. Pilling does not elaborate as to how much of an economy would fit into these categories, but intuitively we can see that it would be large enough to matter. It also seems clear that some countries will be more comparable than others. France and Germany have a lot in common. Electricity can be generated in France and consumed in Germany, and immigration helps keep wages aligned. Per capita GDP comparisons will therefore be quite accurate. Finland and Afghanistan on the other hand have little in common and therefore GDP will be a poorer way of comparing local purchasing power and quality of life.

Lies, Damned Lies, and Statistics

GDP is one number aggregating a whole host of smaller numbers. This makes it useful in many ways, but it obscures what is happening to individuals. For example, an economy could become more prosperous overall while a class of individuals could be worse off. In other words, “the fact than an economy is growing tells you nothing about what is happening to the distribution of wealth,” writes Pilling.

Even worse, people often forget to adjust GDP for the population of a country. If GDP is growing at 3% but the population is growing at 5%, wealth per person is going down. That is why GDP per capita, or per person, is a metric that better reflects how individuals are doing, although it still does not account for inequality.

Finally, GDP is never precisely measured. It is only estimated. Companies, along with individuals are surveyed each year by authorities to gauge economic activity, but a lot of guesswork is used to fill the gaps. Although measurement issues are not too much of a problem in the developed West, they can be crippling in poorer parts of the world like Africa. For example, researchers have had to use satellite images tracking light intensity of the Kenyan countryside to estimate the size of its economy which is otherwise poorly measured by surveys. More strikingly, in 2014, Nigeria finished a thorough revaluation of its economy and concluded that it was in fact 89 percent larger than previously thought!

Incurable Growthism

Perhaps unsurprisingly, despite its many flaws, Pilling does not recommend doing away with GDP. Rather, his critique should be seen as a call to use GDP more cautiously. Firstly, we might wish to adjust it from time to time. For example, the Genuine Progress Indicator or GPI includes the social and environmental costs of GDP. That way, growing GDP by polluting rivers might come out as a net negative and incentivise policymakers to make the polluters bear the cost of their actions. In addition to adding costs, we might also wish to add benefits such as those generated by people looking after their children or parents themselves instead of paying a professional. It is not because it is counted that it does not count.

Secondly, instead of tweaking the measure, we can choose to move away from it entirely. Life expectancy, the number of hours people work, sleep, and rest, literacy rates, unemployment, or self-reported happiness are all interesting metrics that we should pay attention to. None manage to condense as succinctly as GDP the economy, but together their explanatory power is great. We might even wish to look at indices which group these metrics along with GDP. The Human Development Index does just that by combining GDP, educational achievement, and life expectancy together in one number.

I agree with Pilling on both counts. Adjusting GDP or at least being able to know what it excludes is important, and GDP cannot be the alpha and omega. We need a broad range of metrics, some of which might at times give us conflicting information. Where I disagree with Pilling however is in his fundamental belief that GDP growth does in fact tell us how fast our economy is growing. As I will explore in the second part of this essay, I believe that the fundamental theory behind GDP is flawed and that although it is not entirely useless, it cannot measure how much an economy ‘grows’ over the long run. In fact, I do not think that talking of an economy ‘growing’ makes sense. I would rather call it ‘change’.

Part II: The Growth Illusion

A Money Problem

Gross Domestic Product is the total monetary value of final goods and services produced in a region during a period. In other words, it is estimated by adding up the prices of all the final goods and services, which excludes intermediary goods. Therefore, before we go any further, I believe it is important that we discuss prices and where they come from. And since it is impossible to discuss prices without discussing money, I will try to kill those two birds with one stone.

As economists widely recognize, money serves three functions. First, it is a medium of exchange, meaning that it is used to facilitate transactions. Second, it is a unit of account, meaning that it provides a common measure for all goods and services being traded. And third, it is a store of value. The first two functions are relatively straight forward, and I therefore want to focus our attention on the third function.

I like to think of money as a ‘network’ that people can join or leave. People enter this network by exchanging their productive work. I cut your hair, you give me money. The reason I accept your paper in exchange for my labor is because I believe that later someone else will exchange their labor for my paper. When I do so, I exit the network. Although it isn’t exactly right, we can think of money as a way of storing the value of your labor until you feel like exchanging it for somebody else’s, which network participants hope to exchange in the future for somebody else’s productive work. The reason it is not exactly right is because while I hold money, I am hoping that the value of the currency does not change drastically. There is no law that says that because I exchanged an hour of work for $10, that same hour of work will still be worth $10 tomorrow. I am taking a risk. But I am willing to put up with it because when I will cut hairs and exchange them for money I do not need to find someone to barter directly with, neither do I need to exchange my production for somebody’s production right now. Money facilitates exchange and stores my work in exchange for risk.

Another way in which I think of money is as a time machine. Someone could sell lemonade in 1950 and use the proceeds in 2010 to buy an iPhone. Granted, that would require squeezing a decent quantity of lemons, but it is possible. The trick, however, is that first you do not know in 1950 that iPhones will exist in sixty years, neither do you know how much your money will be worth in the future. In fact, your money could be worth absolutely nothing by 2010 if people decide to stop using that currency as a medium of exchange. There is nothing that gives money intrinsic value. The value of the network changes constantly and is defined by how much work people are willing to give in exchange for a portion of the network. Note therefore that for money to have any value, it must be finite. A finite amount of work must be stored in a network of a finite amount of money. For dollars to be valuable, just like gold, there must be a limited number of them. The more money there is, the more units are in the network, spreading the value stored in the network thinner and causing every unit of the network to be worth less. A currency anyone can print more of is useless as you are no longer trading work for work.

Let me illustrate the above with an example. In our simple world, there are ten dollars and ten people. Each person can produce one banana every ten days and each person needs to consume one banana every ten days to survive, however the catch is that people cannot consume their own banana. They must get a banana from someone else. In our simple world, barter is also forbidden and only dollars can be used to trade. Additionally, every person produces their banana in a staggered fashion so that person A produces one banana on day 1, person B on day 2, and so on until person J produces his banana on day 10, at which point we cycle back to person A.

Let us first assume that our simple people all wish to eat a banana one day after they have produced their own. In such an example, person A would produce a banana on day 1, sell it for cash, hold onto the cash for a day, and spend it all the next day in exchange for a banana produced by person B. Person B would then do the same. Notice that in this example, the holding period of cash is one day. Everyone in this example will hold onto cash for a day before spending it. Therefore, if our $10 is to be used as currency to mediate these exchanges, only one person needs to hold onto money at any one time. This means that person A can produce one banana for $10, and buy a delicious banana from person B the next day for $10. In this example, the price of a banana would be $10.

Let us now change our assumptions by saying that our simple people want to eat a banana, two days after they have produced their own. In this example, person A would buy not from person B, but from person C, and person B would sell to person D. This means that the holding period of cash is now two days. In this example, two people would be holding onto cash simultaneously. On day 2, for example, person B sells a banana for cash, and person A is simply holding onto cash he got on day 1 and only intends to spend on day 3. The result would be 2 bananas being ‘stored’ on the network, which totals $10. The price of a banana in this example would therefore be $10 divided by two, or $5. Another way of getting this answer would be to say that the total quantity of money needs to be split evenly between two people because two people hold money simultaneously because of the holding period of money increasing to two.

Although I suspect that by now you get the point, let me give one last example assuming people all eat a banana, five days after selling their own. In this case, person A would sell a banana, hold onto cash, person B would sell a banana, hold onto cash, and so on, until person A finally decides to exit the money network by exchanging his stake for a banana, but at that point, persons A, B, C, D, and E would all hold cash and so the $10 will have to be split five ways to allow the transactions to happen, which means that the price of a banana has dropped to $2. If the holding period of cash were 10 days, the price would be $1.

Note that in all three examples, the same quantity of bananas is produced and consumed and that therefore the ‘purchasing power’ of all participants is the same: one banana every ten days. Yet, the price of the banana is vastly different because of differences in the way in which money is used. The longer people hold onto money, or the longer people delay consumption, the more ends up being ‘stored’ in the network, and therefore the lower the price of a single banana. The equation that comes out of this little thought experiment is that,

Price of stuff = money / stuff made over holding period

Instead of flexing the holding period of cash, let us now change the quantity of goods being produced and consumed. If each person can now produce 2 bananas every ten days, we can quickly see that if the holding period of cash is one day, one person will be selling 2 bananas for $10, or $5 per banana. Conversely, if only half a banana is produced and consumed every ten days, half a banana will be sold for $10 (I know that this does not make much sense, but bear with me), meaning that a full banana would cost $20.

The simple conclusion we can draw from this, which we can check by referring to our equation for price outlined above, is that for a given quantity of money, increasing the amount of stuff that is made over the holding period will cause prices to decrease, while decreasing the amount of stuff produced over that time will increase prices. This fits with intuition. The more we can produce, the more we can buy and so things become relatively cheaper. Therefore, bananas today are a lot cheaper than they would have been to our ancestors five hundred years ago.

Additionally, if the quantity of bananas produced remains unchanged, and the quantity of money increases, say to $20, for people producing one banana every 10 days, and holding onto money for a day, then the price of bananas would rise to $20. This phenomenon is known to economists as inflation, and means a ‘general increase in price’. Conversely, if we reduced the quantity of money $5, prices would fall, which is known as deflation.

To summarize, the price of goods is determined by total quantity of money, the amount produced over a given period, and the amount of time we choose to store value on the money network. More money causes inflation. Less money causes deflation. More production causes deflation. Less production causes inflation. And finally, an elongating holding period causes deflation, while a shortening holding period causes inflation.

We can now better define what we meant earlier by holding onto money being a risk. When our simple people sell their banana for money, they do not know for sure how the production of bananas, the quantity of money, and the holding period will change, all of which will impact the purchasing power of the bills they are holding.

Defining GDP and GDP growth

Earlier, we defined GDP as the sum of monetary transactions in a period. Going back to our ‘simple world’, assuming that our holding period is one day, then the price of a banana would be $10 and GDP over one cycle, or ten days, would be ten times $10 (one transaction every day for ten days), or $100. If the holding period were 10 days however, the price would be $1, and GDP would be $10. Notice that if the holding period of money is the same as the recording period for GDP, then each dollar will change hands once over the period, making GDP equal the money supply. Notice also that changing the holding period changes measured GDP, although production is the same. GDP is affected by the holding period.

Expanding the money supply will also boost measured GDP. As we discussed above, increasing the quantity of money to $20, while keeping production at one every ten days, and the holding period at ten days, would lead to a price of $2 and ten transactions, or a GDP of $20. Therefore, doubling the money supply doubles GDP without changing the production of bananas. In both cases, it seems that we can make GDP grow without making people better off by giving them more delicious bananas.

So, let us now start to improve the productivity of our simple world, while keeping the holding period at 10 days, and see what happens to GDP. Instead of one banana every 10 days, we assume that two bananas can be made over that time. As we discussed above, the price of bananas in this example would halve from $1 to 50c, resulting in a total GDP of $10, exactly like we had before.

Interestingly, unlike shortening the holding period or expanding the money supply, which both increase GDP, increasing the quantity produced has no impact on GDP. In other words, the only thing that makes people better off does not grow GDP. How can this be possible since GDP growth is supposed to track neither a changing holding period or quantity of money but rather an expansion in the productive capacity of an economy?

Quite simply, as economists would quickly point out, none of these estimates are ‘adjusted for inflation’, which means that we have not accounted for the change in price of bananas. If we were to adjust for inflation (or deflation), we would essentially strip the impact of changing prices by assuming a ‘constant price’ to isolate the change in quantity produced. For example, when we decreased the holding period, the price of bananas jumped from $1 to $10. Adjusting for inflation would mean using $1 instead of $10 as our reference price. Similarly, in the example where production double, the price halves from $1 to 50c but twenty bananas are traded in total. Adjusted GDP, or ‘real GDP’ as it often called, would multiply not 50c by twenty, but $1, correctly showing that GDP, or production, doubled. GDP before adjusting for inflation is called ‘nominal GDP’, while ‘real GDP’ uses constant prices and strips out the impact of changes in both the holding period of money and the money supply. When economists talk of economic growth, they most likely mean real GDP growth, or the growth in the quantity of goods and services produced. Saying that GDP grew 4 percent this year means that this year we made 4 percent more stuff than last year. Last year we made 100 bananas; this year we made 104; therefore, our economy grew by 4 percent. If in that year the money supply had also increased, nominal GDP would be higher than 4 percent, but adjusting for inflation, we would record a real GDP of 4 percent.

Something else worth exploring is how to think of growth when more than one good is made. Bananas are great, but there is more to life, after all — I hear. Initially, we could add guavas to our simple world and assume that they take just as long to produce as bananas and taste just as delicious, which would lead to their prices being identical. In the real world, there will be thousands of goods and their exchange rates will be determined by a complex interaction of supply and demand, but for now we will keep things relatively simple. In our case, if both banana and guava quantity produced doubles, we can say that real GDP double. But equally, if instead of 10 guavas and 10 bananas, the following year 20 guavas and 10 bananas are produced, we could see that as simply growing from a production of 20 to 30 units. We can do that because one banana is equivalent to one guava, which we know by looking at the price. If, however bananas cost $1 and guavas $2, then growing from 10 guavas and 10 bananas to 20 guavas and 10 bananas, would mean growing real GDP from $30 (10x$1+10x$2) to $50 (10x$1+20x$2). When we therefore say that the overall quantity of an economy grows by a certain percentage, it need not imply that every single item produced increased by that amount. Rather, we use exchange rates between goods (e.g. how many guavas is one banana worth?) to equate cars with haircuts and measure how much, in aggregate, production increased. Going by our example above, our economy could grow while banana production goes down, so long as guava production more than accounts for the gap.

Putting all of this together and returning to the real world, where multiple things change at the same time, the price of stuff on the marketplace could be simultaneously pushed up by the printing of money and down by the increase in quantity produced. Money supply increasing by, say, 5 percent annually, while quantity increases by 3 percent would yield nominal GDP growth of 5 percent and real GDP growth of 3 percent, with 2 percent of inflation in prices if the holding period of money does not change. In conclusion, real GDP growth tracks the increase in quantity of stuff, whereas nominal GDP growth tracks the increase in quantity of money, if the holding period of money remains constant.

According to US statistics, American per capita real GDP doubled between 1977 and 2018. This implies that the average American can now buy twice as much of what she could four decades ago. But does that make sense? Do people now buy twice as many bell-bottoms and 8-bit Apple II? No. Could Americans in 1977 only afford half of the iPhones that we can now? No. Something’s not quite right here.

Growth Illusion

Real GDP growth or ‘economic growth’ is growth in quantity of production, as discussed above. Over a year or a quarter, GDP growth tries to see how much more stuff is coming out of factories and offices. But does that make sense? Over a year, not much changes usually in the way of technology, but over decades, that no longer holds true. Today, the world economy is not producing more of what it produced in 1950. Sure, we still make cars, planes, houses, and so on, but the quality of these products has improved dramatically, and additional products such as computers, smartphones, and new drugs have been invented in the meantime. Over a long period, an economy does not slowly grow its quantity of production, rather it changes what it produces. Some changes simply mean doing the same thing slightly better, which may be viewed as producing more of the same unit. For example, we could measure the price of lighting in $/lumen and see how that changes through time since light is light. But what about a new drug? This is a whole new category that was never present before. Therefore, talking of long-term growth in output seems wrong.

Yet some economists do just that, reaching some strange conclusions. For example, researchers at the University of Groningen in the Netherlands have produced historical estimates of real GDP for most countries on Earth, often going back centuries. By their estimate, Gambians today each produce the same amount as Americans did in 1800. In other words, modern economic theory would equate the product or output of a country where citizens have smartphones, with that of a country in which most people had never heard of electric batteries, which had been invented just that year in Italy by Alessandro Volta.

A similar comparison is made in Pilling’s book:

“Vietnamese people today are at roughly the same economic level as Americans in the 1880s, but they have the same expectations Americans had in the 1980s. In health terms Vietnam has stolen a march of a hundred years.”

By mentioning the differences in life expectancies, Pilling makes a crucial observation. The quality of life of Americans in the 1880s cannot be compared to that of Vietnamese today. Penicillin, vaccines, and other basics of modern Western medicine are widely available in Vietnam today. No matter how much money someone was willing to pay in 1880 to get penicillin, it would not be available until 1923 when it was invented.

Quality, which comes from new discoveries and inventions allowing us to improve existing products or invent brand new ones, stops us from tracking the change of quantity through time that might permit long-term real GDP numbers to make sense. Instead, we exist in an economy that is constantly reinventing itself. Slowly what we make changes, until it can no longer be compared. A country that constantly makes the same number of cars but constantly improves the quality of its cars is improving but will not show that improvement in GDP number, which simply captures the dollar value of transactions.

To say this in economic terms, is to say that there is no exchange rate between a Nokia phone and an iPhone before the iPhone is invented. The iPhone was not part of the opportunity set of consumption from people in 2006. When new ideas are invented, when we learn how to do new things, we alter the very fabric of our reality. We suddenly find ourselves in a brave new world where iPhones are possible. And just as the idea is invented, so is the possibility of an exchange rate. Now we can see how much it costs us to produce an iPhone and how much people are willing to pay for them. Supply and demand work their magic and a price appears. We can then say that an iPhone costs twice as much as a Nokia phone. But that exchange rates only holds once the idea is out of the bag. The consequence of this fact is significant. Earlier, I explained how we could compare guavas and bananas by using their exchange rates, and how this could be used to measure an increase in overall quantity. But the prerequisite for tracking that increase was an exchange rate that existed throughout the period. We already knew how many guavas a banana was worth and so as quantities changed, we could adjust for price and get a real GDP growth figure. That does not hold for an invention. When we invent something new, an exchange rate appears because of competitive forces, but the new exchange rate does not bridge the gap between before and after the invention. Therefore, talking of an increase in production when we add entirely new items to our opportunity set is fallacious.

A simple way to illustrate this is with a story. Imagine the world’s richest man in 1900, sitting by his dying kid. How much would that man be willing to pay to save his daughter’s life? Everything? Probably. The exchange rate for one life-saving dose of medication at that time would have been enormous. And yet, if we assume that ten years after the girl’s death a cure had been discovered, we would note that its price would probably be much lower. Once the cat is out of the bag and the patent expires, the cost of making that cure would be incredibly low. Because in that new world, where the cure has been discovered, its price is the result of supply and demand between people who know how to make the cure. Once everybody knows, I cannot withhold the idea from you and ask that you pay up. If I did, you’d go next door to someone more compliant. Simply, an idea, before being discovered, is invaluable in that no money can ‘buy’ it. Yet as soon as the idea is discovered, it becomes free — intellectual property rights aside.

Let us now integrate this idea of innovation into an economy where only iPhones and BMWs are made. Every year, 1000 smartphones and 100 cars are made. The going exchange rate is 10 iPhones for 1 car. We will not print additional money, nor do we expect the holding period of money to change. If the next year, 1000 iPhones and 100 cars are still made, but now the iPhone has 4G instead of 3G and cars now have an onboard computer, the exchange rate of 10 iPhones for 1 car may not change and so quantity would remain flat along with nominal GDP, and yet the economy is demonstrably producing higher quality products. That is another issue. In addition to new exchange rates constantly appearing as innovations occur, we might get side-tracked by comparing two items, which themselves are changing and therefore lose all comparability over time.

To account for improvements in quality, economists and statisticians have invented several tools, such as ‘hedonic quality adjustments’, which try to quantify how much better smartphones or televisions are this year compared to last year. Yet, these adjustments are nothing but estimates. There is no measurement one can make that will reveal how much better the iPhone X was compared to the iPhone 8. And that is not even the difficult part. How would one quantify brand new inventions like penicillin? How much better is having penicillin compared to not having it? The price will only tell you what the relative worth of an iPhone 3 to iPhone 4 is today, but it does nothing to indicate the value that inventing the iPhone 4 created. Antibiotics are very cheap, yet their benefits are incalculable.

Because quality adjustments, which really are just estimates, are made, economists can construct what appears to be a long-term growth in quantity. It’s the same stuff you see, just a bit better and with new stuff added. And we can account for the new stuff by looking at how much it is compared to the old stuff. And so now if we take these adjustments we can say that Americans have doubled their income between 1977 and 2018. Or we can say that Vietnamese today are as rich as Americans in 1850. But faced with that proposition, we intuitively understand that is wrong. Today, Vietnamese can give their kids cheap a polio inoculation, an affliction which crippled a 20th century United States president. Every time a new idea emerged between 1850 and 2018, a new exchange rate appeared, which improved the quality of our production in an unquantifiable way. Over long enough, these improvements compounded to create a large difference, which economists cannot identify. We know that that Americans today are richer than in 1977, otherwise we would be consuming the same stuff as we did in 1977 and in the same quantity. But the question of “how much richer?” is not one we can answer. We do not have more of the same. We have new things, which are better. But how much? Nobody knows because now that idea exists, it is a commodity. As the wise man once said: if you cannot count what is important, you make what you count important.

Another way to think about this is that by many measures, the richest Americans living in 1900 did not live as ‘good’ a life as middle-class or even lower-class Americans did in 2000. Healthcare got a lot better. We now have air-conditioning, and access to cheap air transport. The richest man in 1900 could not buy a fridge as good as what a middle-class consumer can get today. So, it is one thing saying that today a person earning $20,000 a year is half a well-off as someone earning $40,000 a year because both can go out and buy the same goods, but how would we compare how much better off the person earning $40,000 is if we also arbitrarily decide to forbid this person to buy anything that uses technology that is less than 20 years old?

My contention therefore is that over long periods, the opportunity set of stuff that is produced and consumed changes so much that the very idea that we can adjust for inflation and therefore track simply “quantity increases” is wrong. Long-term economic growth is meaningless. It cannot be measured, and it is not a thing. Economies do not keep increasing their production of the same stuff every year. Rather, every year, the stuff that we make changes. If we wanted to measure the rate at which our economy prospered therefore, we might think of looking at how fast the goods we consume change. Smartphones are new, so are most Internet-based companies. But we cannot put a number on that change. Interestingly, another conclusion we can draw from this is that the more an economy changes, that is to say the more new products are invented, the less it will seem to grow compared to an economy that would focus on churning out as much of the same old junk as possible. This is because the innovative society does not seek to increase the output of existing products. So next time an economist moans that technological innovations do not show up in statistics, rejoice.

I still love you

Yet GDP is not completely useless. If we were to compare two countries today we could draw meaningful conclusions. We could say that one country for example was twice as rich as the other because the output of one country could literally buy twice as much of the same stuff as the other economy. But 2018 Americans are not twice as rich as 1977 Americans because they do not in fact buy twice as much of the same stuff. They could probably buy a lot more than twice the goods that were made in 1978 because most of that technology would now be outdated. So, GDP can be used to compare the relative wealth of nations at a specific moment in time, but it cannot measure the speed at which a country grows its own wealth. Or more accurately it cannot do it over long periods.

Over shorter periods, GDP is useful. For example, if car production along with the production of many other industries plummets over the course of a few months, this very likely indicates a recession, which we would pick up in GDP. Over very short periods a widespread reduction in the production of most things is a good indicator for an economy going in the wrong direction. But over the long run, the reduction in production of flip phones is not a sign of something going wrong but rather a sign of new and better products, namely smartphones, replacing older categories. One way to summarize this would be by saying that GDP growth can measure us going backward, but not forward. We could measure that half of what was produced last year is produced this year. But as soon as we invent new things, everything breaks down.

GDP growth can also be a true measure of ‘growth’ if for example an economy adds people. If there are more people, we can expect them to make more stuff — more of the same stuff that is. Basically, if the opportunity set of consumption and production is the same, GDP can be used to measure change or compare two sets. My issue is with long periods only.

So, what is a long period? I do not know — a handful of year perhaps. It will depend. The more innovation happens, the shorter it will be since the opportunity set will become dissimilar faster. When the iPhone came out, there was a rupture in the opportunity set of consumers and so comparing anything before that to anything after that would not make sense. Over a year, GDP growth can reliably tell you how much more of the stuff you were making last year you are making now, but it will never tell you how valuable new inventions are, which ultimately is most of what matters over decades. Also note that unless an economy is adding man hours, an increase in output means that something is done differently. In that sense, the very notion of growing the output derives from changing our processes.

Bringing these points together, we could, for example, say that in 2000, the average Chinese earned nearly forty times less than the average American, whereas in 2018, that ratio had reduced to a bit more than six times. In other words, the Chinese got richer faster than the Americans over that period. That much is undeniable. But we have no idea at what rate both were going in absolute terms. In fact, that is not something that makes sense. We need to stop describing our economies as ‘growing’ over the long term. They are not.

What economists refer to as ‘economic growth’ is the increase in the money supply minus some estimate of the increase of price of stuff, but since stuff changes constantly, economists and statisticians must constantly make up numbers using fallacious logic, which ignores that new exchange rates come into existence with new ideas. Ultimately, real GDP growth is not something that can be observed. It is not like height, weight, or heat. Rather, GDP growth is something that we made up by measuring something observable (i.e. nominal GDP growth) and adjusting it with something made up (inflation). While we were manipulating these numbers with complex formulae, we forgot that the whole enterprise made no sense.