Forum Replies Created
-
Replies
-
K/E would give us the price to earnings ratio. I think Blair used the inverse formula: r = E/K
Thanks for the correction, Max.
Lesson: don’t ignore the seemingly obvious!
Shouldn’t we expect capitalists to already include growing perpetuity of earnings when pricing these kind of assets? If so, Blair’s ‘effective’ discount rate actually corresponds to [r-g]. And if the projected growth rate of earnings is 4%, than the ‘actual’ discount rate [r] is around 12% (instead of 8%). Or am I confusing something?
Good point, Max.
Let’s hope I don’t screw up the algebra:
1. In his paper, Blair used the simple formula r = K / earnings — i.e., without earning growth.
2. Had he used the formula with compounded earning growth, we would have:
K/E = (1+g)/(r-g) = 0.068
Rearranging the equation, we get:
r = (1 + g(1+0.068))/0.068
If g = 0.04, we get r = 0.153, or 15.3%
Are you able to share the underlying data, especially the historical data for r and g (or at least point us to it)?
1. Compounded earnings growth g for the past 150 years is calculated from Robert Shiller’s data.
2. Blair Fix computes the discount rate r for all Compustat firms in the U.S. for 1950-2017, based on the ratio of current profit to current capitalization (i.e. assuming future profits won’t grow), as 6.8%. Since unknown future profits are considered more ‘risky’ than known current profits, we set the discount rate, tentatively, at 8%.
I knew a physics professor who, when asked to explain an unclear statement, answered: ‘as I was saying’.
So:
1.
Perhaps I am completely misunderstanding CasP theory. Does CasP assume that the “price is always right,” that share price is always and everywhere determined by the discounting of future earnings, even when there is evidence to the contrary? If so, this would imply that capitalization isn’t really a ritual of capitalists, but an assumption of CasP. Does CasP assume that all capitalists agree that the price is right? If so, why would anyone ever sell? More importantly, why buy a stock that doesn’t issue dividends, which is the case for at least 60% of publicly traded companies, unless you expect the price to increase? Every stock sale has a seller, who is shorting the stock, and a buyer who is going long, i.e., every stock sale involves a disagreement about the future value of the stock.
Yes, the questions you ask here seem to suggest that you might misunderstand our theory. No, CasP doesn’t suggest that the price ‘is always right’ or vice versa, simply because nobody knows what the ‘right price’ is. No, CasP doesn’t claim that share prices are always determined by discounting. There are technical analysts, momentum traders, and other magicians that use alternative rituals, but most if not all believe that, in the long run, prices discount future earnings. No, CasP doesn’t suggests that capitalists agree with one another. And no, CasP doesn’t consider dividends as necessary relevant for stock pricing. As I was saying.
2.
Did they suspend the transitive property of equality without telling me about it?
Transitivity holds when an equation is true. But when an equation is a theory that can be false, you cannot use it to define its own determinants. Concretely in this case, you don’t need to — and cannot — use capitalization to define hype. It is already defined without it. As I was saying.
3.
Our hype debate here is theoretical. The squabble about the S&P500 methodology is practical. Even if the S&P500 isn’t appropriate — and I’m not sure it isn’t — you can devise one that is appropriate. As I was saying.
Perhaps we should agree to disagree.
Some clarifications.
Definition
In our work, we defined:
(1) hype = expected earnings / actual future earnings
Hype has no moral connotations. You can call it ‘forecasting error’, ‘delusion’, ‘stupidity’, or ‘manipulations’.
In retrospect, hype can be measured for any period — a day, two years (as we have done), a century, infinity. If you can specify expected earnings and measure actual earnings in retrospect, hype is their ratio.
Scot is correct that, strictly speaking, hype is impossible to measure all the way to infinity. We cannot stand at the end of time and look back to see previous earnings. However, under normal circumstances, we can offer a good estimate of till-end-of-time earnings and therefore for till-end-of-time hype. The reason is that the discount factor makes more distant earnings less significant for capitalization, so usually 50 or 100 years, depending on the discount factor, will give you 80-90% or more of the capitalization of earnings all the way to infinity. And 100 years of earnings is certainly something we can examine in retrospect.
You can ignore hype if you wish — but then you’ll be able to say nothing about the gap between actual and expected earnings, and about how this gap is being manipulated and leveraged in capitalism for power ends.
Not a definition
In CasP, we theorize that:
(2) capitalization = future earnings x hype / discount rate
(Note that this is a simple expression, and that under different assumptions regarding the pattern of earning, the formula might differ.)
Scot uses this equation to define hype, such that:
(3) hype = capitalization x discount rate / future earnings
In my view, this is a logical error. Equation (2) isn’t true by definition. It is a theory (in this case, of capitalization), and since a theory can be wrong, it cannot be used to define one of its own determinants (in this case, hype).
S&P500: price and EPS
The price index of the S&P500 is total market capitalization/number of shares.
The EPS of the S&P500 is total earnings /number of shares.
Both are free-float measures, and in that sense they weigh larger firms more heavily than smaller ones.
Thank you Scot and Pieter.
1. Can hype be observed/measured?
Hype = EE / future earnings
We can observe expected earnings EE by asking those who expect them what they expect. In retrospect, we can know what the future earnings (E) turned out to be. By dividing one by the other, we get Hype.
In practice, this computations can be complicated because we have to ask everyone who expects and collect eventual earnings data, but these are practical rather than conceptual difficulties. Backward measuring ‘earnings all the way to eternity’ is of course impossible since eternity is never reached, but earnings over the past 100 or 200 years can be measured — and that should do, since discounting reduces the significance of more distant observations.
So no, hype isn’t like utils and SNALT. It can be measured, unambiguously, in principle; and, with enough resources, it can also be measured in practice.
2. Is hype important?
We think it is. The example we give in Figure 11.2 of our book shows that the two-year earnings projections of analysts contain very large errors, and that these errors tend to be systematic and herd-like. Think of what the errors might be for 10-, 50-, or 100-year projections. The key point is that these large ‘errors’ are often leveraged and indeed creordered by dominant capital, and in premeditated ways, and they often have significant consequences for society (think of the roaring twenties turning into the Great Depression and WWII, or of the dot-com and subprime crises having set the stage for the current global turmoil.)
3. What do we know about the power underpinnings/consequences of hype?
Right now, scarcely anything, so it is hard to determine how important it is and in what ways. But that is always the case with new, unexplored ideas. The only way to know is to research them.
More on the re-distributional effect of inflation by firm size
These figures are taken from my 1992 PhD dissertation, Inflation as Restructuring (pp. 419-20). The data pertain to the U.S. manufacturing sector. Each panel shows the experience of a different size-group of corporations: (1) firms with less than $5 million in assets, (2) $5 to $10 million, (3) $10 to $25 million, (4) $25 to $50 million, (5) $50 to $100 million, (6) $100 to $250 million, (7) $250 to $1 billion, and (8) over $1 billion.
All panels show the same relationship: the quarterly net profit markup (net profit/sales) on the vertical axis, and the quarterly rate of producer price inflation on the horizontal axis.
Overall, the panels demonstrate two things:
(1) In manufacturing, the net profit markup is positively correlated with inflation: inflation redistributes income from workers/lenders/suppliers/tax authorities to firms.
(2) This inflationary redistribution tends to become both steeper and tighter with firm size: the larger the firm, the greater and more pronounced the effect of inflation on its net profit margins.
By Shimshon Bichler & Jonathan Nitzan
If you thought stagflation (stagnation + inflation) is an anomaly, think again.
This chart plots the U.S. annual rate of inflation, measured by the rate of change of the implicit GDP deflator, on the vertical axis, and the growth rate of ‘real’ GDP (in constant prices) on the horizontal axis. To eliminate the ‘noise’ associated with both measures, each is smoothed as a 20-year trailing average (that is, each observation on the chart shows the average rates of inflation and growth over the previous 20 years).
The conventional creed
Now, the usual story you’d hear from economists is that inflation and growth tend to go hand in hand: when growth is rapid, the buildup of ‘bottle necks’ causes prices to rise faster; and conversely — when growth is sluggish or when there is a recession, inflation decelerates or even turns into deflation.
With this logic, you’d expect the red line drawn through the observation to be positively sloped. In other words, you’d expect it to start from the lower left side of the figure and rise toward to the upper right.
The actual world
But that’s not what you see in the chart. In fact, you see the exact opposite. For two centuries now, the relationship between U.S. growth and inflation hasn’t been positive, but negative!
On the lower right part of the chart we get ‘growthflation’: a booming economy with low inflation (or even deflation). On the upper left part, we get ‘stagflation’: a stagnating economy with rapid inflation. In other words, we get what economists say is possible only by fluke or due to exogenous shocks.
Yes, that’s right. From the economist viewpoint, the last 200 years U.S. inflation and growth were one big exogenous fluke.
The other possibility is that economists simply have no idea what they’re talking about.
Inflation is always and everywhere a re-distributional phenomenon
In our own CasP research, we’ve argued that this ‘anomalous’ U.S. reality is no anomaly at all.
Capital is power; power means being able to raise prices faster than others; raising prices faster than others requires the threat and exercise of strategic sabotage; and one of the most important forms of strategic sabotage is unemployment and stagnation.
In the capitalist mode of power, stagflation isn’t the exception, it’s the norm.
Further readings
Inflation as Restructuring (1992)
The Global Political Economy of Israel (2002: Ch. 4)
Capital as Power (2009, Ch. 16)
This chart, taken from Blair Fix’s post ‘The Truth About Inflation’ (November 24, 2021), is remarkable for three reasons.
1. It shows the actual variability of U.S. consumer price changes over time — something that I haven’t seen done before.
2. It demonstrates that this variability tends to be far larger than the measured rate of inflation from which it deviates.
3. It confirms that the variability isn’t constant — it changes over time and in ways that are seemingly unrelated to the level/direction of inflation.
On a superficial level, these findings suggest that reported inflation is ‘inaccurate’ — but then, to say that inflation is inaccurate is to totally miss the point. An inaccurate measure is one that deviates from some true magnitude — yet, inflation has no true magnitude to start with.
Commodity prices are set by those who own, sell and buy them; as such, prices reflect relations of power; and power is inherently differential, and therefore impossible to aggregate.
And yet, this is exactly what the measured rate of inflation tries to do — to aggregate relations of power. No wonder this aggregation borders on the meaningless.
In Capital as Power we argued that dominant capital can profit differentially in two ways: by (1) raising the size of its organization, measured in employees, faster than others (differential breadth); and/or (2) by raising its profit per employee faster than others (differential depth).
1. This chart measures differential depth in the United States. It takes the ‘Compustat 500’ — the 500 largest firms listed in the U.S. ranked by market capitalization — as a proxy for dominant capital; it measures the overall net profit per employee in this group; and it calculates the ratio of this measure relative to the net profit per employee in the business sector as a whole.
2. The chart then shows that, in the United States, differential depth has been tightly correlated with the process of inflation, measured by the annual % change of the wholesale price index. The Pearson Correlation between the two series is +0.77 for 1950-1985 and 0.64 for 1985-2020 (excluding the two outlayers of 2008-9).
In other words, when it comes to the relationship between dominant capital and the rest of the business sector, inflation is highly redistributional: it raises the profits per employee of larger firms faster (or lowers them more slowly) than those of smaller firms, and it does so systematically.
(For a similar relationship between inflation and the differential markup of dominant capital, see here.)
Without getting into the details just yet, because I’ve still got a lot of reading to do on the subject, I want to propose here that Regulatory legislation falls into a few very broad categories
The categories you list are potentially useful and informative, but they are not always easy to delineate in practice. ‘Regulation’, like every other important form of capitalist creordering, has multiple consequences, which, in principle, can affect every one of your categories. In order to make this classification useful, we need not only to theorize it, but also to research its multiple consequences.
In the first chapter of our 2009 book, ‘Capital as Power’, we wrote a bit about the limits our journey:
The study of capital as power does not, and cannot, provide a general theory of society. Capitalization is the language of dominant capital. It embodies the beliefs, desires and fears of the ruling capitalist class. It tells us how this group views the world, how it imposes its will on society, how it tries to mechanize human beings. It is the architecture of capitalist power. This architecture, though, tells us very little about the human beings who are subjected to its power. Of course, we observe their ‘behaviour’, their ‘reaction’ to capitalist threats, their ‘choice’ of capitalist temptations. Yet we know close to nothing about their consciousness, awareness, thoughts, intentions, imagination and aspirations. To paraphrase Cornelius Castoriadis, humanity is like a ‘magma’ to us, a smooth surface that moves and shifts. Most of the time its movements are fairly predictable. But under the surface lurk autonomous qualities and energies. The language of capitalist power can neither describe nor comprehend these qualities and energies. It knows nothing about their magnitude and potential. It can never anticipate when and how they will erupt. (19-20)
We then went on to explain why this was the case.
In our view, anything that is conditioned and creordered by the logic of capital as power — including capitalists of all types, workers, corporations, labour unions, governments, NGOs, international organizations, crime networks, etc. — can be theorized and researched by CasP.
But things that lie outside this logic — particularly alternatives to it — cannot be.
The line separating these two categories of course is open to debate.
Which brings me to a question. Differential accumulation is central to CasP, but how do we compare the relative power of states using fundamentally different modes of power? North Korea seems to have gained a position of relative stasis of power through it’s isolationism, and while other states gain and lose power relative to each other, North Korea seems to plod along without significant change.
Good question.
All capital is power, but not all power is capital. The capitalist mode of power is fundamentally different from other modes of power in that it offers a universal measure of capitalized power. As the ritual of differential capitalization spreads, influences, absorbs and incorporates more and more aspects of power, the measurements of capitalized power become more meaningful, encompassing and easier to assess.
This encompassing process started in the bourgs of the late middle ages and continues today. The capitalized aspects of social power within capitalists societies expand, while capitalism itself expands and gradually takes over previously non-capitalist societies (see our 2010 ‘Notes on the State of Capital’).
However, relations of power that do not get capitalized — for instance, the North Korea regime — lack universal quantities and therefore remain difficult to ‘compare’ to other relations of power, including capitalized ones. Moreover, the ‘autocatalytic sprawl’ of capitalized power analyzed by Ulf Martin suggests that the very process of capitalizing power generates its own negation in the form of new uncapitalized power/counter-power, and since this sprawl is self-generating, it implies that capitalized power can never become fully encompassing.
I wonder if you would be able to share the underlying data.
Enclosed is our Excel data file for active corporations reported by the U.S. IRS. The file includes the number of corporations and their business receipts.
Could you clarify how you calculate the markup of the Compustat 500?
Regarding the profit markup for the Compustat 500 and the U.S. business sector: both are computed by dividing aggregate net profit for all firms in the group by aggregate sales for all firms in the group.
