Tuesday, April 23, 2013

The New Way to Measure GDP

A couple people have asked about the new revisions being made in terms of how we measure GDP.  I investigated this and let me just say


I had to go through this monster of a document to find the answer and see if;

1.  the changes are economically and "accountingly" accurate/correct/called-for/legitimate
2.  determine if the impetus for these changes was malicious or well-intended

and after much debating, theorizing and thinking through I think I have your answer.

In short, the changes made (specifically to the R&D) are incorrect in that they double count R&D into the GDP.  However, I do not believe these changes were malicious, but were rather logical and theoretical errors made on the part of the economist and accountant "theoreticians."

The longer version...well....why don't you pour yourself a martini first and find a comfy chair before I delve into it?

You back?

Good, OK, let's begin.

First, I found the subsection of the study that specifically addresses R&D and I have it posted here:

most of this is gobbledygook so let me translate this into English for you.

The red section is basically the UN saying that the value of R&D and all of its future economic benefits should be considered in the SNA or "new GDP measure."  This assertion is true.  The benefits that are derived from R&D SHOULD be included in any measure of economic production for it DOES affect economic production. So so far we have no disagreement with the economists at the UN.

However, we get in into trouble with the green highlighted area.  Here the UN is calling for a way to go about achieving their goal of accounting for or measuring R&D in the new GDP measure.  They don't go into details, merely suggesting new methodologies and techniques to measure this, but here is where they're wrong.

R&D is already accounted for and measured currently in GDP.

The reason why is that all expenditures on R&D are transacted and accounted for.  Maybe not obviously, but every bit of economic production and benefit of R&D DOES end up in the current measure of GDP.

Let me explain why.

When a company, government or other institution spends money on R&D it spends it on two things:

Materials and labor.

Any money spent on materials is immediately accounted for in GDP in that the suppliers of said materials record this as sales.  So just because Apple is developing a new lithium battery doesn't mean all the money it spent on lithium doesn't make it into GDP.  It does via the lithium company producing and selling Apple its lithium. 

That is your first instance of double accounting.

Labor, however, is a bit more ambiguous and the double accounting is not as apparent.  All the money spent on scientists, support staff, and maintenance staff in the R&D lab does not immediately result in the production of a good or service.  Not recordable at that moment anyway.  At this point, however, you may be tempted to make the argument that:

"Well, those people will take their salaries and buys stuff, those transactions of which would end up in GDP."

and think you have found the "double accounting error." 

But, while true, it's Keynesian.  It doesn't account for the fact the value of their labor and expertise in developing a "new lithium battery" is still not accounted for.  For example the scientist who takes his money and buys the Hogan's Heroes DVD set did not MAKE the Hogan's Heroes DVD set.  That is the production of somebody else.  Ergo the labor he expended at the lab is still not making it into GDP.

Inevitably though, the scientist's labor WILL make it into our current measure of GDP in the future when the product or service his labor was used to develop is sold on the free market.

This is an important point to explore further because I believe it is here where the UN economists are goofing up.

Two things happen when this product is sold.

One, a transaction.  This means the transaction will be recorded in GDP.

Two, the scientist's labor is finally valued.  It has a price tag put on it.

The reason why is that when a product or service is sold, you are now attaching a value in terms of dollars AND making a transaction.  So when somebody buys the latest Apple device the price they pay is NOT just for the materials, and NOT just the labor that went into assembling it, but ALSO for the labor in the R&D lab that was expended developing it's latest doo-dad or feature.  Naturally, this is why people are willing to pay a higher price for an IPhone 63 vs. an IPhone 47.  That premium IS the market value of the R&D AND IS ACCOUNTED FOR.

Now, admittedly I did not read through all 800 pages of the UN's SNA publication.  And I could have misinterpreted the writing, reason or rationale of the economists who wrote it.  But as it stands right now, my understanding of this new measure of GDP is flawed and will overstate the level of genuine economic production and wealth in the country.

I VERY MUCH WELCOME any criticisms or corrections from any economists who are more specialized than I am in accounting theory, NIPA accounts, etc. etc.


NormalGuy said...

Meh, just measure production, i.e. manufacturing and resource extraction; and agriculture.

The service industry should be left out, it only recycles money between people anyway.

GDP is not a reliable indicator any longer. It's merely a figure that shows the amount that a nation consumes. It does not show what a nation produces, which is what really matters.

stevierayv said...

I seen this the other day what's the value for unicorn farts those are worth something right

Anonymous said...

While I'm certainly no GDP accounting maven, the concept of using expected future value of current R&D spending (discounted back? and if so at what rate?) as an inclusion for what is supposed to be a relatively simple current calculation (i.e., C+I+G) seems totally off. What are they going to do if the R&D in question doesn't pan out at the expected rate of return? Are they going to backward revise out the sunk cost or use it as a deduction to GDP once written off?

In other words, I totally agree with you. C+I+G already picks up the R&D spending/investment in one or more categories.

If I was a cynical man, which I am, then I would suspect that governments would change this to get more of an GDP accounting boost out of Keynsian government "investment". But our leaders would never try to deceive us would they?

On a side note, have you ever thought of doing a few stories on the content of the Fed's balance sheet and the hidden risks therein? This is something that few people are focused on but as many people as possible need to understand (it's not pretty). Let me know if you need sources.

Jack Schitz

Aeoli Pera said...

No criticism, but I would like to point out a plausible reason for this: incentive for research and, by extension, education because these humanist retards worship that stuff.

Anonymous said...

Dear Cappy,

For those of us who don't have a formal economic education ;-) can you clarify this:

Let's say we have $10 changing hands like so:
- factory worker gets paid $10, and buy sandwich from food stand
- food stand owner takes the $10 and buys meat from farmer
- farmer takes the $10 and buys gasoline for his farming equipment
- Exxon takes the $10 and hires research engineer to look for new wells

(I threw the last one in just for fun! :-)

How many times would the $10 be counted in GDP?

If more than once, would it be right to be confused and angry because it seems absurd? :-)
Shoudln't the GDP be based on some kind of Value-Added figure rather than the crude sale figure?

Prisoner of Gravity said...


Peter Schiff weighs in on the way the GDP is calculated and points out potential reasons why.

James Wolfe said...

This sounds just as accurate and credible as the way govt calculates inflation by adding in rainbows and skittles, and how they reduce the unemployment numbers by putting everyone on the govt dole.

Unknown said...

Peter Schiff talked about this on his YouTube channel, Cap and had slightly different conclusions with regards to the intention, but still worth a look. I just hope people don't confuse his investment strategy with Austrian economics or libertarianism itself, which I feel are three different things.


There have been more developments in my college situation if you saw my email today.

Dance...dance to the radio said...

"this new measure of GDP is flawed and will overstate the level of genuine economic production and wealth in the country"
so they can understate debt to gdp ratios

Anonymous said...

It's the same old (nineties) discussion of 'on/off' balance sheet nonsense. Academics who were educated in the nineties have been exposed to bs to be taught that 'share holder value' trumps everything and is The Measure Of All Measures and that by extension Old Skool Balance Sheets are sooo wrong. Because the intangibles are not on there. Ergo all kinds of distorted and tortuously intellectual efforts to blow up the balance sheet to make 'the assets' more approximate 'Ze Valu' from the (idol worship music) 'Stock Market Value'.

Anyway, this 'pop R&D in there' effort is a similar belated fart from someone - likely now in his forties - academia fart at the UN. The guy has no clue indeed. He is double counting.

Apart from that, any common sense (and people who HAVEN'T been molested by what counts for academia seem to have it in more supply) makes it clear that air = air and the wtf-itude etc.

The CanadianCPA

Unknown said...

I agree with your analysis of when/how R&D gets included in GDP. I think however you may have understated the potential for error in the new method.

I've read Paragraph 10.103 about 12 times, and it's a mess. In one sentence they suggest that the future value of expected benefits needs to be determine, latter they say that Unless market value is viewed directly, it may be viewed as a sum of costs (well duh).

So - how's a government bureaucrat going to see this? My fear is that they are going to try to assign some value to future benefits of R&D in fields they are clueless about (which would be most everything except dispensing red tape). For example, what do you suppose they may view the future benefit of R&D (well R anyway) in Global Warming? It's going to save humanity right - so we can just assign any damn value we want.

At which point they'll be able to point to industry contribution to GDP (SNA) and tell voters how fantastic the spending on Global Warming has been for the economy. (Replace Global Warming with Feminist Studies if that works or you).

Then again maybe I'm just being paranoid.

Unknown said...

"How many times would the $10 be counted in GDP?"

The $10 is counted each time its spent and this is normal and appropriate. GDP is a measure of total economic activity so each time the $10 is spent economic activity occurs to the tune of $10.


Unknown said...

"How many times would the $10 be counted in GDP?"

Each time the $10 changes hands then GDP goes up by $10. GDP is meant to be a measure of total economic activity within a specific time period so this is appropriate.

Anonymous said...

@ Jack: "Each time the $10 changes hands then GDP goes up by $10"

Thanks Jack, but my uneducated self has trouble getting this ;-)

Say a mining company mines $10 of iron ore.
A steel company buys the $10 of ore and sells $20 of steel.
A car part company buys the $20 of steel and sells $40 of car parts.
Honda buys the $40 of car parts and sells a Honda Accord.

In the GDP: wouldn't the iron ore be counted 4 times, the steel 3 times, the car parts twice ?

It seems we should be counting the "value added" at each step, rather than retail price. Value added being the cost of good sold minus the cost of goods and services needed to build that good that were provided by external companies.

Aurochs said...

Soviet style accounting. They want the "numbers" to get better (gdp, u3, ism, home sales, whatever)and the numbers won't for reasons we already understand so they change the formula like they did with cpi. Protect the cathedral mentality. Fuck the truth, fuck rationality, and fuck you. Hold on to power as long as they can - getting and retaining power was their prime motivation all along.
(btw, i think the numbers have always been a fraud. gdp is not something that can be measured like a pint of beer or a batting average. They're just worse now; the lies are right in front of your face.)

Unknown said...

"In the GDP: wouldn't the iron ore be counted 4 times, the steel 3 times, the car parts twice"

I'm not trying to be flippant but that is why it's "GROSS Domestic Product" and not something like "NET ...".

It's a standard number for a specific period (i.e., a quarter or a year) so as long as its apples to apples it's useful.

Cappy's point (if I understand him correctly) is that you are adding in expected FUTURE value of something done today into a calculation that is supposed to represent economic activity that occurs over the specified period. So we're no longer using what happens in the specified period as the measure.

So lets use an example. Let's say I spend $100 doing R&D today and I expect that R&D to pay off at a 200% rate of return in exactly a year (and my cost of funds is 10%). Currently I the $100 I spend today will count to this year's GDP (supplies, wages, rent,...). If I understand Cappy's point (and the little blurb he included), under the new rules not only would that $100 be counted, but what I expect to earn on the R&D would as well so GDP would increase by $300 ($100 of R&D spending + $300 expected return). Or, if we are discounting back the expected return, $280 which is the $100 in R&D + $180 (200 discounted back at 10% cost of funds).

Their are two problems with this new calculation: 1. the return on the R&D occurs in a different time period so we are no longer comparing apples to apples in GDP, 2. the return is conditional upon the R&D producing a return of 200% (what if we are wrong here).

Hope this helps.

Anonymous said...

I'm no expert in this, but pre-IPO biotech companies CANNOT include the "future value" of their patents in a financial statement.

That's what makes these companies so hard to value. Investors have to guess at the future value, because it's in the FUTURE, and therefore does not exist.

They do that by making educated guesses at the value relevance of non-financial information contained in financial statements, which are required by GAAP rules to state only what exists NOW.

Using expected future values to calculate GDP looks like an excellent way to build a house of cards. Which is all fine and well until the cat brushes by and sends the whole thing tumbling down.

kurt9 said...

Anything that involves the UN involves parasitism.

Anonymous said...

They're just trying to be consistent with other forms of investment. GDP has always included BOTH the cost of building a home as investment AND the rental value of housing (imputed in case housing is rental occupied) as consumption, even though the latter pays for the former. It has always included BOTH the cost of building a factory as investment AND the price of the final good as consumption, even though the latter pays for the former. The reason is that a house or a factory is counted as an asset. If you think that knowledge generated by R&D is also an asset, then this revision makes sense.

What's important to remember is that GDP is NOT supposed to measure just consumption; it measures all production, both what's invested and what's consumed.

Here's an example. Suppose Apple invests $100 this year. If it pans out, it creates $500 value next year. If it doesn't, it becomes worthless next year. The revision will change the following:

This year
- Investment: +100 (from 0 to 100)
- Net national income: +100 (from 0 to 100)
- GDP: +100 (from investment)

Next year if project succeeds
- Net national income: -100 (400 instead of 500)
- Consumption: no change (500 either way)
- GDP: no change

Next year if project fails
- Net national income: -100 (-100 due to depreciation instead of 0)
- Consumption: no change (0 either way)
- GDP: no change

So the revision increases I and Y go up the year of the R&D expenditure, and never changes C. Its effect on net national income is +100 in the first year (since the R&D is counted as an asset), but -100 in the second year (since the asset depreciates); it's zero when all the years are added. In an economy (or even a large firm), investment and depreciation occur simultaneously on different things. So in any given year, the change in net national income will only reflect the *change* (investment minus depreciation) in the value of the *stock* (as opposed to the flow) of R&D - this should usually be small.

As you can see, if you focus on net income or on consumption, this revision won't pose much of a problem. And you should really do that, since conceptually, almost *anything* could be counted as an investment. The burrito I bought is an asset for the 30 seconds between the sale and when I sit down to eat it. So you could inflate GDP by arguing that I first invested, and then (30 seconds later) consumed. However, that doesn't change the facts that I consumed one burrito or that Chipotle got $7 from me.

Another thing is that this will not affect GDP growth rates much: it's a one-off bump to the level of GDP.