I: Critical weaknesses in Andrew Kliman’s argument on the falling rate of profit

by Jehu

I have been looking at the numbers in this very interesting paper by Carchedi: “The Law Of The Tendential Fall In The Rate Of Profit As A Theory Of Crises”. I decided to test his 12 reasons why the falling rate of profit is the best explanation for the financial crisis of 2008 — not because I disagree, but because I think he makes a poor case for it. I thought I might go over his material in order to get an idea of how far off his calculations are when measured against a commodity money as Marx used in Capital. The problem, however, is that, so far as I can tell, Carchedi has not published the source data he used to arrive at his conclusions in the paper. (It is possible he published it in a different place, but I just haven’t come across it yet.)

Fortunately, an argument similar to Carchedi’s is also made by Andrew Kliman in his 2011 book, ‘The Failure of Capitalist Production’. And, unlike Carchedi, Kliman generously published a comprehensive spreadsheet of his source material compiled while writing his book. So, with this post, I am going to kick off a comprehensive review of Kliman’s empirical argument for the falling rate of profit thesis.

In the first part of this review, I want to construct a good base line for comparison of the data Kliman provides with his own results. I will be using his original data but employing a commodity money (I will be using gold) to show how commodity money differs from fiat dollars in analysis. After that I will reconstruct his effort on the falling rate of profit thesis again employing gold as the standard of prices.

My justification for this approach is simple: In Capital and throughout his career, Marx employed a commodity money in his analysis to reach his conclusions. I have yet to find a single credible argument against gold as standard of prices in the analysis of empirical data. Thus, it does not have to be demonstrated why a commodity money is necessary for analysis within labor theory. Rather, labor theorists must explain why a commodity money is not necessary — why, in other words, a valueless fiat currency can serve in place of gold as standard of prices.

FWIW: My original contact with Kliman was on this very issue. A blogger, skepoet, introduced me to him because, the two of us thought, using gold might help Kliman in his argument on the falling rate of profit. The blogger’s original post can be found here.

Little did either of us know at the time, Kliman is beyond help. His response to my inquiry was absurdly Kliman-esque (for want of a better term).

“Why is this about “Kliman”? Who else computes or reports GDP or rates of profit in terms of gold? Why single out Kliman?”

No one was singling Kliman out — certainly not me. I just thought at the time his argument could be helped by using gold not dollars. Similarly, this present effort is not intended to be a ‘refutation’ of Kliman’s book, his empirical data or his argument. I am using Kliman’s material because he published a comprehensive spreadsheet with his work. Kliman was so thorough in his approach, he makes it possible to show why using dollars in labor theory analysis may be a terrible mistake.

At the outset, I want to offer two pieces of evidence for why my above argument has analytical validity. I will do this by examining two pieces of data provided by Kliman: Gross Domestic Product and total compensation. I am will also raise critical questions about labor theorists routine employment of common bourgeois price deflators like the Consumer Price Index and Personal Consumption Expenditure index.

When I am done, it should be clear to any objective reader why labor theorists should not be basing their analysis on state supplied data uncritically.

Nominal GDP according to the figures published by Kliman:

First up is Kliman’s data on nominal GDP for the years 1929 to 2007:


And here is the same data as measured using an ounce of gold (exchange value) as the standard of dollar prices:


The second chart was created using Kliman’s own data, but measured in ounces of gold rather than fiat dollars. To create it (and the charts that follow) I simply used the inverse of the price of gold as the standard of prices. If the “price” of gold is $35, the standard of prices (the value represented by one dollar) is 35 dollars in nominal price equals one troy ounce of gold.

However, as you probably already know, over the period under discussion, the “price” of gold fluctuated wildly from year to year, especially after 1971. Thus, over the entire period the standard of prices constantly shifted reflecting a number of different influences beside labor time. I am assuming that on average, in the course of a year, these fluctuations canceled themselves out. The price standard employed in the charts will be the inverse of the annual average “price” of gold.

Note the distinction between nominal measures of GDP and the same GDP using gold as the standard of prices. So far as I can tell, no labor theorist has argued that the value of total United States output fell over most of the 1970s, rose from about 1980 to 2000, and has been falling since 2001. This is because labor theorists look only at dollar denominated measures of output — which shows only a steady rise in GDP — not the actual exchange value of this output.

This is a big error because when might we expected total output (GDP) to fall? During a depression — a crisis — of course. The charts suggest the 1970s was a period of protracted crisis, just as has often been argued by labor theorists. In the data, the only way this depression can be directly seen is when a commodity money is used as measure of value.

And why would this be true?

Because, unlike most labor theorists including Kliman, Marx insisted in Capital that value can only be manifested as exchange value. Similarly, it is only possible to visualize the devaluation of capital in a crisis employing a commodity money — nominal currency measures of GDP cannot show this devaluation. If you are not using a commodity money, you can only speculate there was a crisis in the 1970s — only a commodity money like gold reveals whether there was a crisis or not and how deep that crisis was.

What, perhaps, is more significant for our own situation today, gold indicates we have been going through another depression since 2001. In the six or seven years leading to the so-called financial crisis, the US economy was passing through a deep and protracted depression. Pretty much no labor theorist caught this depression in their data or even realizes it is happening right now. And this means the so-called financial crisis of 2008 occurred within a larger protracted depression since 2001. The near-collapse of the world financial system in 2008, was triggered by this bigger crisis.

Labor theorists have been looking at the financial crisis as an event in a vacuum — an event that has to be explained by its own causes. This has led to all sorts of very odd arguments regarding the rate of profit as the cause of crises; with some very notable labor theorists even arguing every crisis has its own cause. The data as measured by gold suggests the financial crisis was simply a phase in a much larger event that has been occurring since 2001.

Kliman data on employee compensation:

Here below is a chart representing Kliman’s data on employee compensation between 1947 and 2007:


And here is what gold indicated was the exchange value of this compensation during the same period:


As can be seen in the two charts, the data supplied by Kliman looks far different when gold is employed as the standard of prices. Employee compensation fell for most of the 1970s, then rose through the 1980s and 1990s, before falling again in the present depression.

In his dispute with certain other labor theorists, Kliman has argued that the pay workers receive for their labor power has not fallen. Kliman argues a distinction must be made between wages and compensation. Thus, for Kliman, the value of labor is the total compensation the workers receive for their labor power. According to Kliman in a 2011 talk he gave on his findings, the focus on money wages is too narrow:

“It leaves out the Social Security and Medicare taxes that employers pay, retirement benefits some of them pay, etc. All this is part of employees’ ―total compensation.‖ Since the U.S. population is getting older and living longer after retirement, and since health-care costs are rising especially quickly, these additional components of total compensation have increased faster than wage and salary income. In effect, workers are drawing less of their total compensation now, and saving more of it for when they’re older.

Also, the government pays a lot of ―social benefits‖: Social Security, Medicare, unemployment insurance, welfare, veterans’ benefits, etc. The net social benefits (benefits minus the tax contributions that partly pay for them) have increased much faster than wage and salaries.

When all this is factored in, we see that employees’ share of the country’s income has not fallen since 1970, and it’s a good deal higher than in 1960.”

This argument is of no concern to me in itself, except that Kliman uses it as a factor in his argument on the rate of profit and at first appears to have a very big impact on the empirical data. My own preference is to avoid the question of compensation until the category has been taken apart and thoroughly examined critically.

Using the exchange value of the average hourly wage, I can clearly show the average daily wage has fallen over the period since 1970:


But there is a problem here: Kliman presented his data on compensation in aggregate form, while I presented mine as an average daily wage per worker. So to get a better view I had to reissue my data as an aggregate the way Kliman did:

averagedailywage (gold) 1964-2007

I created this aggregate simply by taking my average daily wage data and multiplying it times Kliman’s labor force numbers. In the aggregate form, the data more closely resembles the data Kliman presents, although the increase after 1980 is nowhere near as sharp as his.

In any case, however, as can be seen when comparing my two charts, (average versus aggregate daily wage), all of the apparent spike in wages and compensation after 1980 can be accounted for by the increase in the labor force itself — individual wages and compensation never recovered. Between 1980 and 2001, according to figures supplied by Kliman, the labor forces grew from 99.3 million to 137 million workers – an almost 40 percent increase. Over the period from 1970 to 2001, the average daily wage actually fell, while the labor force increased by 74%, from 79 to 137 million. Thus, in 2001, almost twice as many workers were dividing among themselves the same value as in 1970.

This forced me to take another look at the original compensation figures supplied by Kliman. Having already established that compensation fell in the 1970s, when measured by a commodity money, I wanted to know if compensation recovered after 1980 on a per worker basis. In fact, when measured in exchange value (commodity money) this is not true:

comp_per_person_gold 1947-2007

If the chart above is correct, on a per worker basis, Kliman’s data on compensation, when measured by commodity money, fell sharply during the 1970s and did not recover after 1980. The chart was created by taking Kliman’s data on total compensation, applying gold to it as standard of price and then dividing the result by Kliman’s labor force data. According to commodity money then, total compensation as employed by Kliman rose until 1970, fell sharply throughout the 1970s and never fully recovered after 1980 — just like GDP.

Unless I am missing something, commodity money pretty much kills Kliman’s thesis that compensation has not fallen since 1970 and was higher in 2007 than it was in 1960. Even by his own compensation measure, the exchange value of labor power fell. The exchange value of labor power fell whether this is measured by simple wages or Kliman’s broader definition of “compensation”.

PCE and CPI versus commodity money

I think Kliman makes his error on compensation not out of malice, but because he accepts the inflation adjustments of the fascist state as a true measure of depreciation of the purchasing power of dollars. Everyone knows nominal prices do not convey the true value of commodities and that fiat currency depreciates in purchasing power over time: a dollar in 1990 does not have the same purchasing power as a dollar in 2010. But how do we correct for this when analyzing state issued data?

The two forms of inflation adjustment provided by the fascist state to measure the depreciation of fiat currency that Kliman refers to in his paper are PCE (personal consumption expenditures) and CPI (consumer price index). I have reason to suspect these measures of the depreciating purchasing power of currency should not be used by labor theorists.

First, both of these inflation measures are not measures of exchange value at all but of utility, following bourgeois simpleton economics. In the case of the CPI, a bourgeois simpleton asks: how many dollars it take to get a basket of goods with certain ‘utility’. This appears to make sense, until the simpletons begin substituting hamburger for steak, or a subcompact car for a sedan, etc. Second, it is well known that CPI is not a very good measure even by standard of bourgeois economic theory. It is incomplete and misses vital components of the economy like the entire public sector. Finally, it is highly political and subject to every sort of influence from political forces having nothing whatsoever to do with the actual purchasing power of a currency.

The Bureau of Labor statistics puts this disclaimer on its website:

“The CPI frequently is called a cost-of-living index, but it differs in important ways from a complete cost-of-living measure. BLS has for some time used a cost-of-living framework in making practical decisions about questions that arise in constructing the CPI. A cost-of-living index is a conceptual measurement goal, however, not a straightforward alternative to the CPI. A cost-of-living index would measure changes over time in the amount that consumers need to spend to reach a certain utility level or standard of living. Both the CPI and a cost-of-living index would reflect changes in the prices of goods and services, such as food and clothing that are directly purchased in the marketplace; but a complete cost-of-living index would go beyond this to also take into account changes in other governmental or environmental factors that affect consumers’ well-being. It is very difficult to determine the proper treatment of public goods, such as safety and education, and other broad concerns, such as health, water quality, and crime that would constitute a complete cost-of-living framework.”

Kliman ends up unknowingly — and uncritically — importing this sort of purely subjective estimation of the use value of commodities into his labor theory analysis uncritically.

The personal consumption expenditure index takes the CPI one step further. According to the Wikipedia:

“In comparison to the headline United States Consumer Price Index, which uses one set of expenditure weights for several years, this index uses a Fisher Price Index, which uses expenditure data from both the current period and the preceding period. Also, the PCEPI uses a chained index which compares one quarter’s price to the last quarter’s instead of choosing a fixed base. This price index method assumes that the consumer has made allowances for changes in relative prices. That is to say, they have substituted from goods whose prices are rising to goods whose prices are stable or falling.”

Basically, PCE incorporates the belief the working class will always switch to cheaper substitutes in their consumption when nominal prices rise.

As far as I am concerned any use of measure like CPI and PCE are deliberately designed to understate the extent to which the dollar has depreciated in purchasing power since it was detached from gold. There is nothing of fascist state economic data that can be employed uncritically in labor theory analysis. And there is no measure of the value represented in a fiat currency but how much commodity money its can purchase in the market.

In my next post, I will look at several other categories Kliman includes in his rate of profit argument to establish a base-line for those as well.