GDP and its critics

I've just finished reading Diane Coyle's excellent book on the history of GDP. For a measure that has only been used since the 1940s, GDP has amassed an extraordinary following. Whole economic theories now depend on it. 

But there are those who argue that GDP is fundamentally flawed and should be replaced. Sadly, some of them show a lamentable lack of understanding of accounting and the methods used to calculate GDP. For example, here is Steve Forbes:
GDP represents the value of all final products and services. It ignores all the steps that go into the making of these things. It’s sort of like looking at a carton of milk and paying no heed to everything that goes into creating that milk and getting the carton onto the store shelf.
GDP thus gives a distorted picture of the economy. How many times do we read that consumption represents 70% of the economy and therefore it’s important to “stimulate demand” by increasing government spending?
And he goes on to recommend using gross output (GO) instead of GDP, arguing that it "measures the economy in a far more comprehensive and accurate manner".

To show why this is seriously flawed, I need to explain a bit about how GDP is calculated. Experts on this will no doubt call me out on being far too simplistic - GDP is an incredibly complex measure and its calculation is close to being a black art. But this post is not aimed at experts. It's aimed at ordinary people whose opinions are swayed by the likes of Steve Forbes. 

GDP can be calculated in three different ways - the expenditure approach, the income approach and the production approach. The first of these adds up all the forms of expenditure in the economy, giving us the familiar formula:

GDP = C + I + G + (X-M)

where C is consumption spending, I is investment spending, G is government spending and (X-M) is net exports(imports).

Steve Forbes complains that imports are treated as negative for GDP. But this is a no-brainer. If you buy imports, money leaves the country. If you sell exports, money comes into the country. Purchases of imports are therefore correctly a negative for GDP. (But profits from sales of products made with imported goods are positive.)

The second way of calculating GDP adds up all the sources of income in the economy:

GDP = W + P + R + I + D+ (T-S)

where W = wages & salaries, P = corporate profits (or operating surplus), R = rents, I = interest, D = depreciation and (T-S) = net taxes after subsidies.

As total spending = total income, the first and second methods should give approximately the same result, though measurement errors do create some difference. 

But it is the third way of calculating GDP - the production approach - that Steve Forbes is complaining about. It takes gross output (yes, THAT gross output) and eliminates intermediate outputs to give the gross value-added output for the economy. Like many, Steve Forbes does not seem to understand why intermediate outputs are eliminated. But it is because intermediate outputs are used to create final output. To show this, I'm going to use Steve Forbes' own example of milk production. (Please note the figures are entirely fictitious. I have absolutely no idea how much it actually costs to produce, package and market a litre of milk.) 
  • Farmer X has a herd of dairy cows. The cost of maintaining his dairy herd is $1,000. They produce 1,000 litres of milk which he sells to a supermarket for $1.10 per litre. Gross sales are $1,100 and profit is $100.
  • Manufacturer Y produces the cartons for milk. Production cost is $200 for 1,000 cartons and he sells the cartons to the supermarket for 25 cents per carton. Gross sales are $250 and his profit is $50.
  • The supermarket puts the milk into one-litre cartons and sells it to customers at £1.80 per litre. The supermarket's unit costs are $1.10 + $0.25 = $1.35 per litre and they have staff costs and overheads amounting to a further 15 cents per litre. So if they sell all the milk, the gross sales revenue is $1800 and profit is $300.
The gross output is the total sales revenue from all three of these, i.e. $1,100 + $250 + $1,800 = $3,150. But there's a problem. Since the supermarket has to pay for its supplies, the sales revenue of the suppliers is already included in the supermarket's production cost. So by adding in the suppliers' sales revenue, we are double counting. This is the same problem that accountants face when creating group accounts: if you include sales from a subsidiary to its parent, you artificially inflate the sales revenue of the whole group. It's known as grossing-up. Here we are talking about supply chains, but for national accounting purposes the same applies. If you include intermediate outputs, you gross up the private sector's value-added contribution. To obtain an accurate figure for value-added GDP, therefore, we need to eliminate the intermediate outputs. When you do this, it is evident that only the supermarket's sales contribute to GDP. The rest disappear. 

So GDP does not recognise the value of intermediate outputs. But that doesn't make GDP wrong, any more than a set of consolidated group accounts is wrong because it has eliminated intra-group transfers. What Steve Forbes is proposing is replacement of consolidated business output with grossed-up business output to make the contribution of business to the national economy look bigger. I think they call this "cooking the books". 

Currently, the US's National Income and Product Accounts only show gross output by sector. It is now planning to produce overall gross output figures as well, in parallel with GDP figures (this is the change that Steve Forbes is crowing about). This is a good addition to the national accounts: we do need to understand the contribution of intermediate producers, and sector analysis isn't always adequate since final products may be made from intermediate sources in several sectors. Also, in these days of global supply chains, intermediate sources may not be in the same country as the final output, in which case the intermediates arguably should be regarded as imports and their cost deducted from gross output (no wonder Steve Forbes wants imports to be treated as positive!). So gross output will be a useful measure. But it is not by any stretch of the imagination a replacement for GDP. 

But it is the conclusion that Steve Forbes draws from this that worries me. Returning to the familiar expenditure equation GDP = C+ I+ G+ (X-M), Steve Forbes in effect argues that grossing-up business output would considerably inflate the importance of I in relation to C and G, since all those intermediate outputs would have to be included in investment spending. This would, in his words, "put business and investment in their rightfully large place". But this is bizarre. In what way is milk bought daily from a farmer "investment spending"? It is simply a cost of production and therefore belongs in C, not I. If you are grossing up, then businesses consume too.  

Of course, it suits businessmen like Steve Forbes to claim that the driver of the economy is business output and consumption is an optional extra. But the sole purpose of business output is consumption. Businesses don't produce products unless there are people willing and able to consume them. Without C, there would be no I.

And this brings me to Steve Forbes' view that government spending is a negative. In days gone by, when governments mainly raised money to finance wars, this was a reasonable view: money spent on wanton destruction leaves the country never to return. But that's not what the majority of government spending is used for these days. Most government spending goes into the economy: whether or not it is effectively spent entirely misses the point. Businesses sell to government just as they sell to households. Government spending should therefore correctly be seen as positive in the GDP expenditure calculation, just as household consumption and business investment are. 

So Steve Forbes's criticism of GDP is ill-informed and irrational. But GDP is certainly not without its problems. Its extreme complexity makes constant revision inevitable, and even with all that complexity it is still at best only a rough indicator of economic growth. If the inputs to the GDP calculation change significantly, the result can be large swings in apparent economic standing that have nothing to do with reality - for example, Nigeria has just reported an 89% increase in GDP that is entirely due to rebasing the components of GDP. And it leaves out all sorts of things - such as unpaid "home work" - for no particularly good reason other than that they are difficult to measure. Critics of GDP complain that it doesn't measure wellbeing (true), it doesn't adequately account for intangibles (also true), it gives greater importance to "making stuff" than providing services (true, but not surprising given when it was invented), and it ignores depletion of natural resources. All of these are valid criticisms. 

Coyle gives a good summary of these criticisms in her book. But she then dismisses most of them. Replacing GDP, or changing it to include other factors, is not the point. The real problem is that we are expecting far too much of GDP. There cannot be "one measure to rule them all". The fact that GDP focuses narrowly on measures of expenditure, income and output is not a bug, it's a feature. We need other measures as well. Coyle suggests a "dashboard" of economic indicators for policymakers to enable them to judge the health and prospects of the economy and set policy accordingly. GDP (nominal and real) would be included in this dashboard but would not be the sole driver of monetary or fiscal policy. 

As I am no fan of single targets such as CPI or NGDP, I like Coyle's idea. I am encouraged by the fact that both the Bank of England and the Fed are now using a "basket" of indicators, though they have not yet ended the primacy of the inflation target, because we still have not exorcised the 1970s inflation demons. But I am hopeful that we will gradually move towards a more balanced approach and start to include other indicators such as wellbeing and sustainable resource use in our measures of economic health. 


  1. Some very good points. There's just unnecessary criticisms of GDP.

    A good example is found in the book Understanding National Accounts here:

    Page 16 onward: example on pasta and flour making. Almost same as yours.

  2. I have long had doubts about GDP and its workings and what it has been used for. Grand Delusional Potential was something that came to mind.

  3. TravisV here from themoneyillusion comments section.

    The market monetarist view: targeting inflation < targeting NGDP < targeting wages.

    I hope you agree that it makes more sense to target incomes rather than inflation. And while targeting wages is preferable to targeting NGDP, it would be politically infeasible. So the market monetarists settle for targeting NGDP.

    Also note that Glasner, who is more left-wing than Sumner, agrees that targeting inflation < targeting NGDP < targeting wages. Glasner's inspiration is Ralph Hawtrey, who advocated stabilizing the wage for unskilled labor.

    Here are some posts from Glasner and Sumner discussing targeting wages vs. targeting NGDP:

  4. "But this is a no-brainer. If you buy imports, money leaves the country. If you sell exports, money comes into the country."

    I think about this in a slightly different way. Imports get consumed, so they show up as a positive number in C, but they were to produced in the domestic economy, so they need to be subtracted in (X - M).

    Similarly, exports were produced in the domestic economy, but they will be consumed in some foreign economy, so they got produced, butt don't show up in C, so they have to be added in (X - M).

  5. "but they were to produced in the domestic economy".

    Sigh. That should be "but they were not produced in the domestic economy"

  6. Haha. And "butt don't show up in C" conjures up some interesting images too.....

  7. You say "The gross output is the profit that all three of these make, i.e. $100 + $50 + $300 = $450" But isn't gross output the sum of the market value of goods and services, not profit? There's no double counting when you sum profit - this sum, after all, is part of the income calculation, isn't it?

  8. Hi Frances

    Your 'milk' example doesn't ring true to me.

    You say "The gross output is the profit that all three of these make, i.e. $100 + $50 + $300 = $450." But I thought that output related to sales, not profits.

    In this case, the gross output would be the sales of all three - $1,100 + $250 + $1,800 = $3,150. But GDP - the total "added value" at each stage - is $1,100 (milk) + $250 (cartons) + $450 (filling & selling cartons) = $1,800.

    It's easy to see that the end product of all this - what the country is capable of making - is 1,000 filled cartons of milk, and those sold for $1,800.

    That means that GDP ($1,800) makes sense, and Gross Output ($3,150) is an over-estimate.

    Sorry if I have the wrong end of the stick!


  9. David and Alex,

    Yes, you are correct. My bad. Should be sales, not profits. I will correct.

  10. Under the income approach wouldn't it be more accurate to also include capital gains income? Capital gains is also value added just like on a newly produced good.

  11. It's obviously difficult to sum up the state of an entire economy in one figure alone. But still, GDP gives a pretty good idea of what is going on.

  12. Thank you for sharing valuable information. Nice post. I enjoyed reading this post. The whole blog is very nice found some good stuff and good information here Thanks.Also visit my page Investment Property Accounting Rental Property Accountants New Zealand.

  13. GDP does say something but not everything. your need to calculate the GNP by adding the exports and subtracting the imports to get the best view of your countries position.

  14. FC,

    In the milk example, I don't think there is an issue with double counting, because the whole process, from farmer to supermarket shelf isn't done by the same entity. The farmer, container producer, and supermarket are each separate entities with its own costs and its own, different final product. The milk, container, and the contained milk are all final products in and of themselves. For example, farmer could have sold that milk to anyone, other supermarkets, individuals, bakeries, etc.

    'In what way is milk bought daily from a farmer "investment spending?"

    It depends on what is done with it. Individuals using it for their breakfast are consuming the milk. A local bakery is making an investment. Even the supermarket sales of milk don't necessarily constitute the end of the road as that milk sold could be involved in some other productive process. It's the same with all goods, the computer I'm using to write this could have been used to run some internet business, but it was bought by me for surfing the internet and other leisure activities.

    GDP is not 'wrong,' but what it isn't what it purports to be. It doesn't measure 'economic growth,' but merely how much was spent in the economy in a given year. Any accurate representation of economic growth must include a complete analysis of the entire productive process, from the natural resources to the supermarket shelf. Ignoring everything that happens before the supermarket phase is like pretending that the goods just appear in the supermarket. As was correctly stated in the Forbes article 'consumer spending is largely the effect, not the cause, of prosperity.'

    'But the sole purpose of business output is consumption. Businesses don't produce products unless there are people willing and able to consume them.'

    If you want to break it down to a sole purpose, it would be to profit. Yet profit can only happen if people are willing to consume them at a higher price than your costs, as you've said. However, this willingness is anticipated by the producer. There is no way to know for sure if the willingness is genuine until the products are produced and the registers ring, which is where the risk lies. The order of actual, real life events is as followed:

    Producer anticipates a need/want > Producers Invest in labor/capital needed for production > Production > Goods brought to market > consumption

    Consumption is the last step in the process, totally reliant on every step that came before, and thus cannot drive anything. Thus any measure of economic output that ignores all but the last step is going to be severely lacking. That's not to say it has no use, but economists have essentially erected monuments to its existence which is a dangerous thing.

    1. I'm sorry but I just don't agree - and I am arguing from a business standpoint, not economics. Businesses don't invest if they don't think there will be sufficient future demand for the product or service to justify the investment. Therefore consumption - or perhaps more accurately the expectation of (future) consumption - DOES drive the economy.Demand is not merely a passive response to supply.

      Having said that, businesses can create markets for their goods and services by clever marketing, which creates demand. So it is perhaps best to recognise that both investment and consumption drive the economy in different ways. We can regard consumption as the "pull" to business investment's "push", if you like.

    2. Oh, and on the milk example. It doesn't matter to whom they sell the milk. The point is that if the milk is used to make yoghurts for sale, rather than simply poured on someone's breakfast cereal, the purchase price of the milk is included in the sale price of the yoghurts. Therefore if you add together the sale price of the milk and the sale price of the yoghurts you are double counting the sale price of the milk (which is also its purchase price).

    3. You're making the common error of taking 'desire' as 'demand' as it is defined in economics. What you described is 'desire,' which does comes first, and that is ultimately what spurs action. However, desire on its own gets us nowhere. The very first, recordable, actionable step in attempting to quantify that desire comes through investment, followed by production of a good or service. Then, and only then can economic demand be exercised.

      Put it another way, I can desire a good all day long until I'm blue in the face. I can have infinite money such that in theory I exercise demand constantly. Yet as long as no one has invested in the required resources to produce the good, I will never, ever be able to consume it. I have never said that 'demand' is a passive response - but it IS a response nonetheless. 'Desire' is the real force here, but there is no sure way to quantify that economically.

      As for the milk example, it does matter who the milk is sold to and what is done with it. Yes, the milk that went into the yoghurt is reflected in its price, but for what you're saying to hold we'd have to assume that ALL of the milk produced went into yoghurt. We have no way of knowing that, we only know that the milk producer made XYZ milk. A portion of that was used for breakfast, and a portion was used as an intermediate good in other products. Within the context of that one example I can see your point, but as it pertains to the GDP discussion it doesn't fit because the individual producers make goods that have many different ending points.

    4. Nope. Future demand, as I explained it, IS a driver of business decisions and therefore of the economy.

      And it IS possible to distinguish between products that are intermediates and products that go to end consumers. You are thinking far too much about suppliers and nowhere near enough about purchasers. Of course it is possible to find out how much milk is bought by the makers of yoghurt. It's in their accounts.

    5. Let me ask you directly: Is there anything wrong with this bit from above?

      'Producer anticipates a need/want > Producers Invest in labor/capital needed for production > Production > Goods brought to market > consumption'

      Producers anticipating a need/want is the same thing as anticipating a future demand for it. But at that EXACT point in time, no money has changed hands, no steps have been taken to satisfy that future demand.

      Acting as though consumption is the start of the process requires you to believe that those goods materialised from thin air (and even then the only difference is that whatever cosmic force doing the materialising takes the place of the 'conventional' producer as the starting point)

      I really don't know why this is so contentious, but you're far from alone, it's a standard position in modern economic discourse

    6. Producers don't "anticipate needs/wants". They go and find out what people want. Why do you think they do focus groups, surveys, market research? It's not just to find out what ALREADY sells. It's to find out what MIGHT sell.

      There are some inventions that do apparently appear "out of the blue" because some genius has dreamt them up and they just happen to be the right product at the right time. But even then it isn't that simple. Do you really think Jethro Tull invented the seed drill because he fancied inventing a seed drill and it happened to be the right product at the right time? Nonsense. He knew how much time and manpower it took to sow crops. He had quite possibly listened to farmers moaning about how long it took and how many men they had to employ to get their crops sown. So he invented something that was a lot more efficient both in time and manpower.

      Sometimes you have to look beyond mere economic theory. Business practice is much more illuminating.

    7. "Producers don't "anticipate needs/wants". They go and find out what people want."

      They do nothing of the sort. What anyone wants other than the so-called 'producers' themselves seek to gain is irrelevant. What matters is the ability of a firm to borrow against its own accounts, from its own lenders: the 'producers' borrow until they cannot any more, they put the entire proceeds into their own pockets. At some point they can no longer borrow ... The firm is emptied out of its own credit, it is a husk, bankrupt and is abandoned ... but this does not matter because there is the next firm, and the next firm, and the next firm and the next = Bain Capital, Elliott Management, etc. Wall Street in general.

      The process has a purpose: someone knows but is not saying there is no such thing as a productive (industrial) enterprise: such things cannot exist, they are thermodynamically impossible.

      There is no real gain to the public or workers from the self-lending process, nor is there any gain from myriad Wall Street Ponzi schemes whereby finance system losses are forced into the economy! These finance system losses are (mis-) tabulated into GDP as 'investments' = currency traps. Financial misrepresentation/fraud amounts to 70% of GDP; the only real productive enterprise is agriculture (which is also underwater due to mis-tabulated/mis-measured input costs). This leaves out ordinary criminal activity such as drug rackets, robbery, arms trafficking, buying and selling politicians and 'regulators' including multi-trillion-dollar money laundering activities on the part of central banks.

      Firms borrowing against their customers' accounts -- whereby the customers' banks create deposits in firms' bank accounts with customers being required to retire these loans with their labor that is discounted against the firms' use of (further) discounted machines, so that the human worker is always 'underwater' relative to his mechanical, debt-subsidized competitor ... whose input (capital) costs are purposefully mispriced to include only acquisition: land-lease-royalty (less subsidy) + unit access-extraction expenses (also misrepresented). See Graeber: human workers are fetched with a choice = wage slavery by discounting their own labor to compete w/ robots or unemployment. The surplus gained this way is a 'positive' component of GDP as well, just like slave labor in antebellum Dixie was good for the planters.

      Honest accounting = thermodynamic whereby a human labor @ btu is equal to the machine potential @ btu, done this way a gallon of gasoline is measured against the btu output of human labor and costs $2,500 ... at this price there is none of the wasting infrastructure because capital (means of production) is too valuable to waste. Now?

      There are converging adverse trends: the aggregating costs of capital annihilation on one hand, the looming costs of not annihilating our capital on the other. Do you stick the gun barrel up your butt or stick it in your mouth, not much of a choice but this is what we have left for ourselves in our drive to enrich a small handful so that we might live our lives through them vicariously on television.

      Stupid humans. We deserve what we get.

  15. one should not conflate GDP and GDI because they are measurements of two different could say they are both ways of measuring the size of the economy and that they usually amount to the same in current dolars, but we cant say that GDP = W + P + R + I + D+ (T-S) because that's the measurement of gross domestic income, and they are calculated separately...

    here's an interactive graph from FRED showing both over 5 years:
    run your cursor across the face of the graph to see their differences quarterly...

    those numbers shown are in current dollars, but for the purpose of measuring the change in GDP, current dollars are pretty useless, because that change could simply be in prices for goods and services being measured....all percentage changes in real GDP are in chained 2009 dollars (or until last year, chained 2005 dollars) to adjust for inflation...what that gives us, in terms of the quarterly reports, is not a measure of the change in dollar value of our goods and services, but a measure of the change in our units of output...that way should the price of milk in your example double when less milk is produced, it does not increase real GDP, but rather decreases it...

    here's a stub from the BEA showing some of the metrics they use to deflate various components to obtain the change in units of product:

    1. Thanks rjs, that's interesting.

      GDP and GDI *should* be the same, since income = expenditure, but measurement errors mean that in practice they are not. I did say that in the piece.

    2. i knew you had that caveat in your post, Frances, but that formula (GDP = W + P + R + I + D+ (T-S) ) had been bothering me since i first saw it two days ago, so i had to come back and make that point...

      i havent given " income = expenditure" enough thought to know if i agree with it or not, but if that's true, yes, then current dollar GDP and GDI should be identical...

      the broader point i wanted to make was that real GDP is not a dollar measurement, as is commonly perceived, but a measure of units of output....i struggled with GDP reports for a few years before i realized that's why they were converting each of its components into those chained dollar amounts....that use of dollar amounts to represent units of "product" is part of the reason GDP is so poorly understood...

    3. rjs,

      I don't agree. Regarding GDP as a measure of output leads to the mistake that Steve Forbes has made. If all we are measuring is output, then eliminating intermediate outputs is wrong. But GDP is not simply a measure of output. It is an attempt to measure the size of the economy in dollar terms at a point in time, using three different measures which should be equivalent - income, expenditure and output. That's why eliminating intermediate outputs is necessary: if you gross up output, it far exceeds income and expenditure.

  16. here's the most recent report:
    in table 3, there are billions of dollars on the left, and billions of chained 2009 dollars on the right..
    the measurement in billions of dollars is the measurement of the size of your economy in current dollars..(often quarterly at an annual rate)
    the measurement in billions of chained 2009 dollars is a dolar denominated proxy for units of output..
    "real GDP", and all the quarterly and annual comparisons throughout the report refer to the later..

  17. Very good summary of different ways to calculate GDP.


    "But this is a no-brainer. If you buy imports, money leaves the country. If you sell exports, money comes into the country. Purchases of imports are therefore correctly a negative for GDP."

    But if somebody had said, when they first put these GDP accounts together, that exports subtract from GDP, and imports add to GDP, the whole accounting system would still have been coherent. It is just a convention that we add exports and subtract imports when somebody invented the GDP system, surely,

    If we export, we have more money, true. But we have less goods.

    If we import, we have more goods, but less money.

    As C + I + G is all goods and services, we should add imports (also goods and services) as well. So I would also agree with Forbes that imports add to GDP. That way GDP would be a much better indicator of wealth on the economy, all goods and services used in one year.

  18. "Steve Forbes complains that imports are treated as negative for GDP. But this is a no-brainer. If you buy imports, money leaves the country. If you sell exports, money comes into the country. Purchases of imports are therefore correctly a negative for GDP. (But profits from sales of products made with imported goods are positive.)"

    I don´t understand why imports should be negative for GDP. In GDP = C + I + G + (X-M) , If M goes up, that means there is also a change in C, I or G. So M is balancing item to avoid that for example an increase in C caused by increased consumption of imports is added to GDP.

  19. Very late - but Larry Elliot today (27th April) gives us a good reminder that many times we use GDP when we should be using GDP per capita...

  20. just want to make a note here that according to revised figures, nominal 1st quarter GDP was at $17,101.3 billion annually, up from the $17,089.6 billion annualized figure of the 4th quarter...
    see table 3:

    but everyone is reporting it as down at a 1.0% rate, because the measurement they're reporting is inflation adjusted, hence a measure of the change in the actual output of goods and services produced by labor and property in the US...


Post a Comment

Popular posts from this blog

WASPI Campaign's legal action is morally wrong


What really happened to Signature Bank NY?