The Fight for the Formula - The Great Invention: The Story of GDP and the Making and Unmaking of the Modern World - Ehsan Masood

The Great Invention: The Story of GDP and the Making and Unmaking of the Modern World - Ehsan Masood (2016)

Chapter 2. The Fight for the Formula

It is not easy for a free community to organise for war. We are not accustomed to listen to experts or prophets. Our strength lies in an ability to improvise. Yet an open mind to untried ideas is also necessary.

—John Maynard Keynes,
How to Pay for the War (1940)

The crash of 2008 and the recession that followed it are often talked of as a second Great Depression. But there’s no comparing our most recent banking crisis with the misery that incinerated lives in the 1930s. Life for many today is difficult, certainly. But the Great Depression was a disaster.

In America at the height of the economic slump between 1932 and 1933, somewhere between one in three and one in four men of working age was out of work. One million families lost their farms and two million people became homeless. Families in jobless, homeless, and indebted circumstances were becoming ill, and disease was leading to death on a horrific scale.

At the same time, after the stock market crash of 1929, banks had either locked their doors or were in the process of slamming them shut to protect their remaining deposits from panic-stricken customers intent on emptying their bank accounts.

The Depression was deep, and in the United States both the federal government and Congress were in disarray over what to do about it. Part of the problem in knowing what to do was an almost complete lack of data on the health of the economy. This is sometimes difficult to appreciate in our age of Big Data, but before GDP the volume of data available to the public was a fraction of what it is today.

The only similarity is that both then and now, governments were desperate to know how economies could be kick-started, and especially how new jobs could be created. But to create new jobs, governments needed to first know how many were employed. To raise living standards, they needed to know what people were earning.

Among the lawmakers most concerned about the absence of meaningful information on the numbers of people in jobs or the value of factory goods was the Republican senator for Wisconsin, Robert La Follette Jr.

La Follette was a centrist Republican in that he championed relief for the unemployed and also supported government spending to create jobs. He was also keen that the federal government should begin compiling statistics and in 1931 presided over the following piece of political theater to prove his point.

At the time, members of the Senate were holding an inquiry into the state of American manufacturing. Congressional inquiries are an opportunity for elected representatives to ask searching questions of those government employees who are paid to carry out Congress’s wishes, but who rarely appear in public. Witnesses are called to testify in what can be a tense and sometimes gladiatorial atmosphere. In 1931 the sacrificial lamb was Frederick Dewhurst, an official in the Department of Commerce who was ordered to give evidence before the US Senate. Judging by the transcript of the exchange,1 Dewhurst probably didn’t know what was about to hit him.

It is worth reproducing something of that exchange:

CHAIRMAN (LA FOLLETTE): I would like to direct your attention to the available labor statistics. What would you say about the statistics of unemployment? Have we any?

DEWHURST: Not any, except for one month in the history of the republic. That was April 1930. The census taken in that month gives a sort of benchmark.

C: And they took a sample test later on in January?

D: A sample in January 1931, of 19 cities only.

C: But aside from that, we have nothing?

D: No, we have no direct measure; we can infer or estimate or guess on the basis of changes in employment.

C: You can go around Robin Hood’s barn to get it, but so far as actual statistics are concerned, we have not any; is that correct?

D: Yes.

C: What about payroll statistics? We are woefully lacking in wage statistics, are we not?

D: Yes.

C: Now let us consider statistics on consumer purchases.

D: In the first place, we do not have a very good coverage of the total field of consumer purchases. In the second place we have lost the identity of commodities completely by the time they get to the consumer, except for a few, such as automobiles.

C: You cannot answer the question.

D: I would hate to guess it within 100 percent.

C: What about statistics on savings and investments?

D: … We do not know much about that except from the published statistics of new security issues.

C: How about investment credit? Do you make the same answer?

D: That is the same answer; yes; except that it is more so.

At this point in the proceedings the transcript shows Democratic Senator Morris Sheppard, one of the founders of the temperance movement, stepping in to protect Dewhurst from further self-harm. Sheppard asks the civil servant whether he thinks that a more complete picture of the US economy would “be of benefit to the government and to the people of the country.” Dewhurst’s reply is telling.

“In my opinion, it would be most desirable,” he says, before adding this coded warning: “May I add that a statistician is prejudiced in that he always wants more statistics. Statisticians are never satisfied.”

La Follette’s plan was to persuade the US government to find out and then publish what people were earning and to establish whether earnings had been increasing or decreasing in the years leading up to 1932. In June of that year he got what he wanted; the Senate passed a resolution ordering the commerce secretary to carry out an estimate of total US national income for 1929, 1930, and 1931 and to provide this information before December 15, 1933.

That task was assigned to Simon Kuznets (1901-1984), who had emigrated to America from Russia at age twenty-two, in 1922, five years after the Russian Revolution. Kuznets started work on producing national income accounts the following month, January 1933, and submitted his final report to the Senate a year later, on January 4, 1934.2

When the Senate asked for an estimate of the value of the US national income, there was no precedent, nor a template for how this should be done. It meant that Kuznets had to invent the method for calculating total national income for an economy the size of the United States. He had to do so in less than a year with a small team and with none of the infrastructure for collecting data that we take for granted today—not a single computer; not even a hand-held electronic calculator.

Kuznets began work by compiling lists of America’s different industrial sectors, such as agriculture, mining, and factories, from where data would need to be found. He was given a small team of three “senior assistants” and five “statistical clerks.” Together they hit the road, visiting factories, mines, and farms, interviewing owners and managers, and writing down figures in notebooks. Once a notebook was filled, it would be sent to the Department of Commerce in Washington to be checked or revised. Books would often be sent back and forth between Kuznets’s team, Washington, and the original source, until anomalies and queries were ironed out.

True to his word, Kuznets delivered on time and the Senate published his work as a 261-page report entitled National Income, 1929-32.3 The report’s price tag was the princely sum of 20 US cents and it turned out to be a surprise best-seller, with its 4,500-copy print run selling out well within the year.4

Kuznets and his team made no effort to varnish their findings and confirmed what was plain to anyone at that time. Between 1929 and 1932, US national income had almost halved. The total income distributed to individuals in 1929 had been $81 billion. In 1930 it dropped to $75.4 billion. The following year it was $63.3 billion, and in 1932 it had crashed to $49 billion, confirming the misery and devastation everyone could see around them.

Moreover, no sector had been immune to the crisis. Agriculture income had dropped from $6.3 billion in 1929 to $3.4 billion in 1932. During the same period, income from construction had collapsed from $3.1 billion to $0.9 billion. The biggest fall had been in manufacturing, from $18.1 billion in 1929 to $8.4 billion three years later.

Salaried staff on the whole suffered comparatively less (their incomes fell from $5.7 billion to $3.4 billion). In contrast, incomes of those on (less secure) daily and weekly wages crashed from $17.2 billion in 1929 to $6.8 billion in 1932.

There was, however, one sector of the economy where incomes not only had stayed constant but had even managed to register a small increase. This was the column marked “Government.” According to Simon Kuznets and his team, between 1929 and 1932 income distributed by government entities was up from $6.5 billion to $6.8 billion, reflecting a small increase in public works programs.5

In many ways, the rise in government spending isn’t as intriguing as is its inclusion in Kuznets’s national accounts. This is because, for much of his professional life, Simon Kuznets fought the inclusion of government spending in GDP. Data from businesses formed the cornerstone of his methodology, and he was confident in the quality of the data he and his team had gathered. Government data was something else. Kuznets believed quite stubbornly that government spending does not belong in the national accounts in part because spending by the state doesn’t make for a very productive economy. When governments spend money, they’re not producing anything new, just recycling what is there. At the same time, he took the view, as do many economists today, that spending by governments crowds out spending by businesses. If governments produce a good or a service, then that is a good or service denied to business, which, ultimately, isn’t good for economies. Such an approach, as we shall see, would put him on a collision course with his US government colleagues in the years ahead.


Although for the next decade Kuznets would be Washington’s go-to man for all things national accounting, a new generation of economists working for the US government had become uneasy about Kuznets’s accounts as being the sum total of private incomes.

By the late 1930s the drums of war were getting louder and leaders in Europe and America felt that conflict with Germany was a matter of when, not if. Many had either fought in or lived through the First World War, and they knew that they needed to be better prepared militarily a second time around. That meant having deeper knowledge of what the government had available to spend to go to war. Handguns, rifles, tanks, battleships, and fighter aircraft would all cost, as would professional soldiers and volunteer conscripts. Businesses, too, would be affected. Factories would need to be commandeered for the war effort, and that would mean making and selling fewer of their normal products.

Herein lay the federal government’s big problem. According to Kuznets’s methodology for economic measurement, not only was there no space for the state’s finances, but its absence would make a wartime economy look even smaller. That is because people would be earning less; because business would be producing less and because the vastness of government spending would be kept off the books. US government economists needed a system of measurement that would go beyond adding up private incomes, as Kuznets had done, and include the role of government.

The problem for this younger generation of economists was that Kuznets was having none of this. The future Nobel laureate resisted his colleagues’ attempts to change his methods and include government spending in the national accounts, so much so that they looked for an alternative champion.

Such a champion would emerge in the shape of intellectual-cum-socialite John Maynard Keynes (1883-1946).6


John Maynard Keynes (known to his friends as Maynard) was certainly a one-off. To many he was the 20th century’s greatest economist, though certainly not an economist in the conventional sense. He was a part of the generation who lived through the end of empire, the Boer War, the Great Depression, and two world wars. He was a polymath, gifted with unusual powers of curiosity and a remarkable ability to influence.

Keynes had that rare gift of being able to persuade. Using a combination of confidence, intellect, charm, dogged persistence, and a certain amount of bullying, he believed he could persuade not just his fellow economists, but also those in disciplines outside economics, and those with no connection to academia at all, to come round to his point of view. Keynes was also a compulsive writer of newspaper articles, an assiduous builder of networks of friends in high places, and an extraordinarily successful investor. His efforts would not always succeed, and many of those he encountered, especially in the United States, found him insufferably pompous. But there is no doubting he knew how to work the corridors of politics, which he would do through sheer force of argument and his not inconsiderable presence. These are among the reasons why the British economic historian Lord Robert Skidelsky, author of a three-volume biography of Keynes, calls him “The Master.”7

Keynes wasn’t born lowly, but neither was he superrich. He came from a family of empire-era scholars that included Charles Darwin before him. These were men born into relative privilege, and they benefited directly from being part of the British imperial experience. But rather than profit from that experience as so many did, they instead chose a path of scholarship and discovery.

So how did Keynes put the G in the GDP equation?

Let’s begin with the book that put Keynes on the world economic and political map. While Darwin is famous for On the Origin of Species, Keynes’s magnum opus is called The General Theory of Employment, Interest and Money. It was a publishing sensation when it appeared in 1936. It is still in print, still much more readable than your average economics text, and, perhaps because of this, still consulted by world leaders in times of economic crisis. It is the primary source for any argument in favor of increasing public spending.

At the time of the Depression, Keynes was also consumed with the question of how to get more people back into jobs and how to help businesses become more productive. Keynes’s prescription was radical for the times, and he called it “demand.” What he meant by “demand” is that if private businesses were unable to create jobs during a severe depression, then governments needed to step in and spend money on big projects, such as roads, bridges, buildings, even towns. These projects in turn would create the kinds of jobs that would last for some time, which would help people get back to work and acquire new skills. Such an idea has become known as a Keynesian stimulus. It was also the framework for President Roosevelt’s New Deal and was applied again to deal with the latest banking and financial crisis. After 2008, across Europe and America, governments poured billions into expensive schemes such as roads, dams, and bridges (or “shovel-ready” projects, in the words of President Barack Obama), to help people get back into employment.8

But the idea of a Keynesian stimulus was—and continues to be—opposed, especially on the center-right of politics. That is because, as Kuznets pointed out, governments on their own are not productive in any economic sense. They merely spend what they receive, and that is largely from the taxes they collect from their citizens. Large government expenditure according to this view risks spending beyond what a country can afford. Also, during Keynes’s time, the other great exemplar of massive government spending to boost an economy happened to be one Adolf Hitler, not a model to be imitated.9

According to the alternative view, the most important factor in creating jobs is what is called “supply.” “Supply” in this context means that if businesses are left to get on with making and selling things, then jobs will follow. Instead of thinking they can help, governments need to do the opposite: they ought to get out of the way and let businesses supply goods and services. That will create demand; it will create jobs and consumer spending, all of which will keep the economy going. On no account, according to a more extreme interpretation of this view, must government provide money for jobs, impose barriers on business, or write new rules or enact new regulations. Recessions and depressions happen when governments try to meddle. Even if they think they’re meddling for the right reasons, the results will invariably be catastrophic, which is why the role of government is simply to stand aside, almost not to exist.

The idea that supply creates its own demand is often attributed to an eighteenth-century French businessman and economist named Jean-Baptiste Say. In fact economists misleadingly call it Say’s Law. I say “misleadingly” because it’s not really a law as this word is understood in the scientific world. Science, too, is just as argumentative as economics. But an idea in science gets to be elevated to legal status only if it can be proved, or verified beyond doubt—essentially when the arguments largely stop. If you kick a football, for example, few will debate its trajectory. It will behave in a way predicted by theory and confirmed through experiment, according to one of Newton’s laws of motion. In contrast, when an economist uses the word “law,” he or she usually means a theory or a hypothesis. It is an idea, yet to be confirmed, and subject to differing opinions. That is very much the case with Say’s Law.

A modern example of the use of Say’s Law could be the rise of technology giants such as Google and Apple. Both these companies have created massively successful products. Because these products did not exist previously, you could say that their supply created demand, which, in turn, has created millions of new jobs around the world. Supporters of Say’s Law claim that these are all jobs in which government has had little or no role. But critics of Say’s Law, such as the economist Mariana Mazzucato, would argue the opposite. They say that much employment creation involves some role for government. The tech giants, for example, wouldn’t have a business were it not for the Internet and the World Wide Web. The Internet was created by scientists funded by the US government; the web was funded almost in its entirety by the governments of Europe through the European Organization for Nuclear Research, CERN, in Geneva.10

Until the Great Depression, Keynes did not fundamentally disagree with Say’s Law, but once employment levels began to fall off a cliff, he began to do something that many scientists find quite difficult: in the face of a changing situation, he started to think differently and to ask questions. Was unemployment rising because businesses were producing things that no one wanted to buy? he asked. Or was it more likely that jobs were being lost because people had become too poor to spend any money in the shops? Was it this lack of demand that meant entire industries were becoming idle, laying off workers in the thousands?

At some point during this period Keynes concluded that he had to change his mind regarding Say’s Law. The problem of unemployment during a depression, he now believed, was not one of supply but one of a lack of effective demand. This is where consumers are in a position to pay for something. It is not demand in the abstract. According to Keynes, one way to stimulate such demand was for governments to spend money and create jobs. Keynes convinced himself and others that the urge to consume pushed economic activity forward and the urge not to consume held economic activity back. In one lecture in 1933 he described his ideas on demand and consumption using an imaginary scientific experiment on chickens.

Imagine, he asked his audience, that scientists could remove that part of a hen’s brain that controls the urge to eat. This hen is completely normal, except that if you put a pile of wheat in front of it, the hen won’t touch it. The hen is so uninterested in eating that it eventually starves to death. Something similar, Keynes concluded, was happening to the world’s economies after the Great Depression. There was no shortage of supply, he said, because businesses had the ability to produce whatever society demanded, whether wheat or chaff. What was killing consumers was that they were unable to consume the existing wheat. The wheat was there. They just couldn’t eat it. It’s similar to the kind of argument used sometimes to explain famine. Hunger doesn’t happen because of a lack of food, but because what food there is doesn’t reach the people who need it the most.


In Chapter 18 of The General Theory, Keynes set out in more detail his arguments for public spending in a depression. He argued that businesses cannot just be left to get on with things because, in a depression, there will be few with the means to purchase. In making his arguments he divided expenditure into two components: spending by households, which he called consumption, and spending by businesses, which he called investment.11

According to Keynes, if consumption increased, if more people were able to buy goods and services, then businesses would spend more. When businesses spent more on goods and services as well as on premises and machinery, then that would lead to an increase in investment. Taken together, an increase in consumption and an increase in investment would cause the national income to increase.

Keynes according to his biographer deliberately chose not to write his idea in mathematical form, but if we did it could be summarized as:

Economic Output = Consumption (C) + Investment (I)

But we need to bear two things in mind: First, Keynes had not yet written his ideas as a table or a formula. And second, although his book provides a rationale for greater government spending, he hadn’t at this point included public spending in his definition of how to measure an economy. That would come next.12


In 1939, when Adolf Hitler invaded Poland, Britain declared war on Germany, and Keynes’s thoughts turned to the same question of war financing that the younger US economists had been asking before him.

“It is not easy for a free community to organise for war. We are not accustomed to listen to experts or prophets,” he mused in a series of articles for the London Times. “Our strength lies in an ability to improvise. Yet an open mind to untried ideas is also necessary,” he would say.13 One such “untried idea” was figuring out the income and outgoings of the British government. As it was for his US counterparts, this was the basic knowledge on which all other thinking around wartime finance needed to be built, and Keynes, like Senator Robert La Follette Jr., was incredulous that this information was not easy to find. “The statistics from which to build up these estimates are very inadequate,” he would say. “Every government since the last war has been unscientific and obscurantist, and has regarded the collection of essential facts as a waste of money.”

In the absence of official estimates, Keynes drew heavily on the work of one of his Cambridge colleagues, the statistician Colin Clark (1905-1989). Clark was an obsessive collector of data on the national income. He was the nearest that Britain had to a Simon Kuznets, although he was a Kuznets with no official mandate.14 “There is no one to-day who does not mainly depend on the brilliant private efforts of Mr Colin Clark,” Keynes would write. “But in the absence of statistics that only a government can collect, he could often do no better than make a brave guess.”15

Colin Clark, just like Kuznets, was mainly interested in collecting data on private incomes. And as with Kuznets there was no role for government in his national income scheme. For Keynes there was no point in studying wartime finance if government spending data couldn’t be found. In the very least he needed such data to work out how much the state could expect in taxation and so arranged for Clark’s tables to be updated.

Keynes calculated the size of the UK economy as of March 31, 1939, to be £4.85 billion. He subdivided this total under three headings:

“The consumption of the public”: £3.71 billion

“Privately-owned capital equipment”: £290 million

“Cost of the services provided by government”: £850 million

He converted his Times articles into a book, and at eighty-eight pages long, How to Pay for the War is one of Keynes’s shortest books. Its main function was to help the government and the public understand the nature of British wartime finances. But there, buried in a table on page 14, was the final variable in our GDP formula—the G, for government spending, which he calculated to be £850 million. In 1940, six years after Simon Kuznets had presented his national income estimates to the Senate, Keynes had written down in a table the basis for what today is the formula for GDP.

GDP = C + I + G

Keynes’s London Times articles and his book How to Pay for the War caught the attention of policy makers. In Britain, it prompted the Treasury to take a closer look at government finances and to have them published for the first time the following year, in 1941. But in America, it arguably had a more powerful effect. Keynes’s book supported those economists who were moving away from Kuznets’s insistence that an economy be measured as the sum of private incomes, and who wanted to find a way to include government spending in the national accounts. They were in the Department of Commerce, as we have seen, but another group was working in the Office of Price Administration and Civilian Supply (OPACS). This was the government agency responsible for mobilizing resources for the war, and they, as much as the Commerce Department, needed a handle on the nation’s finances. But such was Simon Kuznets’s hold on accounting methodology that these younger officials were finding it hard to have an impact.

By 1940 and with America’s role in World War II looming, OPACS desperately needed to know what government spending would look like. They tried to argue that including government spending would likely boost the size of an economy. However, others countered by saying that doing so could also result in a post-war economy shrinking once government spending on armaments, for example, was cut back.

We know that Keynes had a former student at OPACS, and during his 1941 visit to America to negotiate the lend-lease agreements, they met at Keynes’s suite at the Mayflower Hotel and exchanged documents.16 One of those documents was an estimate of US gross national product, which is worked out in a similar way to GDP.17 This exercise gave them an idea about how much America would have to spend on the war.

The problem that OPACS faced was Kuznets, who protested against the new system. This quite possibly prompted Keynes to note, “There is too wide a gap here in Washington between the intellectual outlook of the older people and that of the younger. But I have been greatly struck during my visit by the quality of the younger economists and civil servants in the Administration.”18

But by this time, the momentum was turning against Kuznets as Keynes’s international influence grew. Kuznets began to be criticized in the pages of influential journals, which helped Keynesian economists in the US administration to join the debate and to say that they, too, agreed that Kuznets was wrong to ignore spending by the state, and that the US national accounts had to change to reflect this.


Academics, including statisticians, are detail obsessives—they need to be. The best can function rather like a camera lens: looking at the bigger picture one minute, then zooming in on some tiny but significant detail the next. This is a rare and impressive ability—and both Keynes and Kuznets had it. They also had something else in common. Just as Keynes was uncomfortable about expressing his ideas using algebraic symbols for fear of misinterpretation, so, too, Kuznets’s earliest writings on national income show his worry about the possible misuse of the work he was doing for the government. Kuznets took pains to explain to his readers that his estimates were an estimate of incomes and should not be interpreted as a picture of American economic activity, less still a proxy for economic health.

For the next four decades, Kuznets would complain, writing articles in journals, speaking at conferences. He would argue that, once the war was over, GDP should go back to what it was. He would say that GDP, as currently constructed, was “destined to become one of the most used and misunderstood sources of economic information.”19

Keynes’s relationship to America would always remain complex. He had his critics, including those with whom he was negotiating. Whether during the negotiations over the lend-lease agreements, or later at Bretton Woods, which led to the creation of the World Bank and the International Monetary Fund, anyone on the other side of a conference table experienced the full force of Keynesian pomposity. The British art historian Lord Kenneth Clark, an equally precocious high achiever, recalls a man who did not know how to “dim his headlights.”

Kuznets wasn’t wrong to believe that a measure of incomes is a far better guide to well-being and to what is important to citizens. But he misjudged the US administration’s determination, indeed its insistence, that public spending be included as a positive indicator for economic health. By 1941 Simon Kuznets’s days as America’s chief national accountant were over. It was Keynes who had effectively helped his younger American colleagues win the argument and defined GDP for generations to come.