At the turn of last century, in 1898 and 1908, Rosa Luxemburg writes her pamphlet ‘Reform or Revolution?’ as a response to Eduard Bernstein’s desired reformist platform for the Social Democratic Party. In the second chapter, she attacks Bernstein’s view that capitalism had adapted its institutions and practices as to avoid capitalist crises (namely the regular recessions which usually occurred around every 10 years). Bernstein argued that newly-formed financial systems and monopolies—at the time referred to as ‘the credit system’ and ‘capitalist combines’—had allowed for such coordination among capitalist decision makers to prevent future crises. Bernstein thought Marx’s prediction that capitalist economic crises increasing in severity had been halted by these new institutions of credit and monopolization, which he believed allowed capitalism greater, and more efficient economic coordination.

On the contrary, Luxemburg argued that these credit systems and monopolies were part of a natural progression of capitalist economies towards larger crises. She thought the consolidation of economic power would open the doors for more frequent and more extreme crises. Part of the logic of her argument was that these developments in capitalism would mean larger crises in the eventuality that one of these growing capitalist industries runs up against limitations in resources, or overextends beyond the capacity of the economic system to support their risks.

“The belief that capitalist production could ‘adapt’ itself to exchange presupposes one of two things: either the world market can spread unlimitedly, or on the contrary the development of the productive forces is so fettered that it cannot pass beyond the bounds of the market. The first hypothesis constitutes a material impossibility. The second is rendered just as impossible by the constant technical progress that daily creates new productive forces in all branches.”

—Rosa Luxemburg, 1908
‘Reform or Revolution?’
Chapter 2: The Adaptation of Capital

While Luxemburg was proven more correct than wrong in her Chapter 2 analysis, her scientific starting point felt somewhat sketchy to me, and the conclusions she drew felt too strong for her relatively armchair-theoretical justifications. (To be fair, a considerable amount of contemporary economics largely consists of puffed-up theoretical assumptions.) The empirical evidence that she invoked felt too broad to be able to motivate such specific conclusions, and her arguments struck me as somewhat dogmatic in their Marxism. It felt like she was starting from the orthodox Marxist conclusion that ‘increasingly larger capitalist crises are inevitable’ and had gone in search of social processes that would justify orthodox Marxism in the face of Bernstein’s arguments. That is not to say that Bernstein was being more scientific—in fact, he was likely a more egregious offender of empiricism than Luxemburg.

At our point in history, we can see that a lot of Luxemburg’s general ideas around credit systems and monopolies have been confirmed, while some of her other ideas have not.

She was essentially correct about credit. Credit’s primary functions “to extend production and to facilitate exchange” has indeed allowed for more investment/growth; greater extension of capitalist industries to the point of overextension; and a separation between decentralized stock ownership and centralized company management which allows for more extreme and uninformed forms of speculation and hoarding.

She seems to have been more hit-or-miss on the effect of monopolies, correctly predicting that they will provide cheaper goods and services in developing markets, but apparently incorrectly arguing that they will inevitably break themselves apart as profit margins go down and equalize with other sectors: “In the face of the increased difficulties of finding markets, each individual portion of capital will prefer to take its chances alone. At that time, the large regulating organizations will burst like soap bubbles and give way to aggravated competition.” Instead, it seems that monopolies choose to maintain their large level of internal coordination. Instead of breaking off parts of their capital to compete with each other for profit, they subsidize and support new ventures with the profits and scale of their existing structure.

When I first read this ‘hit-or-miss’ chapter from Luxemburg, therefore, I mostly wrote it off as being an interesting piece of economic history, but not presently useful. I thought it was a nice summary of how a Marxist might interpret finance and monopolies, albeit being empirically wanting and not terribly helpful as a model for how to go about economic research and theorizing.

~

That was how I filed that chapter away in my memory—until the day I stumbled upon ideas from Adam Tooze’s history of the global economic crisis of 2008-2012 in his book Crashed, along with his related lectures and talks. I was struck particularly by the financial/institutional consolidation which came about as a result of the 2008 crisis.

The parallels between the consequences of the global Great Recession and Luxemburg’s analysis of finance and monopolies was rather striking—in both cases, as a response to the largest economic crises of their era, financial institutions were consolidated in an attempt to be a kind of regulating and supporting force behind capital. For Luxemburg, this was the creation of novel credit systems. For us, after the 2008 crisis, Europe’s major banks have become more dependent on U.S. financial institutions than they were before:

“[T]he crisis was not merely American but global and, above all, North Atlantic in its genesis. And in a contentious and problematic way [the American-centric narrative] had the effect of recentering the world financial economy on the United States as the only state capable of meeting the challenge it posed. That capacity is an effect of structure—the United States is the only state that can generate dollars. But it is also a matter of action, of policy choices—positive in the American case, disastrously negative in the case of Europe.”

— Adam Tooze, 2018,
Crashed: How a Decade of Financial Crises Changed the World, Introduction

Alongside the increase in financial consolidation, capitalist monopolies and conglomorates would be able to take advantage of this larger financial capacity to grow to even larger scales. Even though finance expands how much capacity it has to support the vagaries, risks, and speculation around capitalist industries, the capitalist industries would then grow to fill the space. The capacity that market-making institutions have to effectively stabilize capitalist ventures shrinks as monopolies grow. As a result, the capitalist risk-taking and individualist profit-seeking actions once again become a threat to the capitalist system.

The similarities between Luxemburg’s historical situation and ours reveals a potential historical pattern in the capitalist system, which I characterize like this:

1. In response to economic crises, market-making institutions like credit systems consolidate their power a bit more.

2. As a result of this consolidation, capitalist industry grows larger, to the extent that this new financial system can support them.

3. The capitalists then are likely to overextend in an inventive way which was not anticipated (or ignored) by regulators of the markets.

4. The financial system, which has now been streched thin to accommodate the scale of industries, is unable to take the economic hit in its current form.

5. A capitalism-threatening crisis occurs, and in the bourgeois scramble to repair the situation, lawmakers further consolidate and expand finance/market-making-institutions, as to better absorb the hit to the economy.

6. The cycle begins again, with a more consolidated and powerful financial system, but also with the capacity for proportionally larger monopolies and capitalistic opportunism—potentially opening the door to even more significant and interconnected global crises in the future.

This isn’t to say that this large crisis cycle is necessary—there are other factors which might intervene in this admittedly abstract cycle (e.g. limited global resources, hard government control over the economy as a result of regime changes to fascism or socialism). We might discover whether this basic abstract pattern has actually occurred historically, if we had more in-depth research. But this is a blog post, so in lieu of that research, I think it’s worth noting that the pattern I described at least seems plausible. A very charitable presentation of history might go:1

• The late 1800s Long Depression led to the creation of those credit systems and monopolies which were analyzed by Luxemburg and Bernstein.

• Following the speculative and monopolistic Gilded Age, the 1930s Great Depression led to the Keynsian expansion of government influence on financial investments.

• As capitalists took advantage of government spending and investment, the 1970s stagflation crisis broke out. This led to the U.S. dropping the gold standard entirely, and consolidating even more monetary power into the Federal Bank. Developments in the next 40 years would see an increased importance of the U.S. dollar in the global economy, as well as highly multinational corporations and supply chains.

• Subprime mortgages and related speculative consolidations were a get-rich-quick scheme for capitalist investment bankers, which put pressure on a financial system that was spread thin, in a way which the government technocrats didn’t see coming. This led to the global financial crisis kicking off in 2008. The response to the crisis has resulted in further consolidation between the U.S. and European banking systems—Now, Europe is even more deeply and directly dependent on the US Federal Bank compared to before 2008.

There are certainly recessions and smaller crises which occured between these large system-threatening crises. However, the pattern I described focuses on crises which are large enough to genuinely destablilize the capitalist market-making institutions. This threat essentially forces the bourgeoisie to significantly develop and consolidate power in the financial system to create an environment that can thwart the immediate threat to capitalism; as well as an environment which can support an even larger, more globalized market. This financial, market-supporting consolidation allows for capitalist industries to grow to an even larger scale, freed from the restrictions of the previous, more constraining system. In my charitable telling of history, it appears that the institutional solutions to the previous crises each came with their own new loopholes and incentives for the more globally-connected crises to come; these loopholes were not obvious at the time, but more clear in a historical reconstruction of the lead-ups to the crises.

Thoughts on the Marxist approach to economic science...

With this historical pattern, we can return to Luxemburg’s analysis of the consolidation of financial power and the consolidation of capitalist companies. Instead of Luxemburg providing a purely in-the-moment argument about monopolies and finance, her analysis fits into a broader historical pattern regarding these consolidations of economic power. This way of carrying out economic analysis is informative for how we might scientifically study the political economy.

This underlying analysis that I gestured at in this blog post—the dialectic between market-making institutions and the growing capitalist powers which push the market-makers to globalize and consolidate—describes a way of theorizing about economics in big-picture, epochal scales. The scale of analysis which tries to encompass ancient, feudal, and capitalist epochs takes seriously the possibility—even likelihood—of an eventual collapse of capitalism, and encourages our theoretical language to reflect this.

More importantly, it emphasizes a way of understanding how Marxists go about theorizing in very abstract terms, while still trying to be empirically grounded and, ultimately, reliably capable of predicting the outcome of potential political actions.

~

To begin explaining this style of theorizing, I think it is good to emphasize that Marxists predict things at an abstract level. In this case of credit and monopolies, a prediction we might make would be at the level of abstraction of: “Capitalist consolidation, expansion, and individualist profit-seeking will create economic crises that negatively hurt a lot of people who are not part of the plutocracy. These crises will push for further consolidation and coordination of market-making institutions—usually meaning the increased centralization of banks, finance, regulations, and monetary-control. These new institutional forms will tend to support even greater capitalist consolidation, expansion, and individualist profit-seeking tactics—opening the doors to future, more globally interconnected crises.” This is almost a generalization of Luxemburg’s analysis that the institutions meant to adapt capitalism to crises are too weak against the development and growth of capitalism.

Because we predict things at such an abstract level, we aren’t really describing the more detailed mechanisms by which the next crisis will occur. Instead, we talk about things probabalistically and in terms of the material development of the capacities of interacting institutions. For example, “The market-maintaining institutions are slow to adapt, and less likely to proactively anticipate coming crises. Consolidation of market-making institutions often means fewer redundant mechanisms to protect against individualist actions by profit-seeking capitalists. On the other hand, capitalist industries are driven by the structure of the market and internal developments to grow, monopolize, and seek out new profit-making ventures. They grow their economic power to the degree where they once again push up against the boundaries of the market created and supported by the market-maintaining institutions. Now that companies have grown to match the scale and coordination of the market-making institutions, the capitalist’s short-term profit-seeking ventures will increasingly come at the long-term cost of weakening the stability of the markets. The more this occurs, the more likely one of these ventures will be the final straw that triggers a full-blown economic crisis.” As these institutions develop over time, in the way that we understand them to here, it becomes more likely that a crisis will occur—the proactive capitalists outpace and overextend beyond the regulative capacities of the sluggish and reactive market-maintaining, market-making institutions.

This abstractness can lead people to accuse Marxists of being able to explain away any challenge to their theories. They say that essentially any new information that comes in can be described in the Marxist model, because the Marxist model is so vague and abstract that it allows for this. The problem here is that the Marxist model should be judged at the level of abstractness which it predicts. Often, Marxist theories also exist across multiple layers of abstraction. Layers can be proven wrong, and will need reworking and added nuance. However, a more abstract descriptive layer might be minimally affected; while the more concrete levels might become more complex. The overarching Marxist theory thus develops in a piecemeal fashion.

To be more scientifically rigorous, we would want to be able to examine our abstract theories at a more precise and concrete level of abstraction. We would want to examine the exact legal codes and flows of capital. We would want to see what areas capitalist industries were expanding to, and what strategies they use to maximize profits. Particularly, we would want to look at how exactly companies act in profit-seeking ways, outside of the spirit of the market-maintaining laws—laws which the economic system generally relies on to remain stable.

That sort of specific research can help us understand things at a more concrete level. That is, where the next crises are likely to come from, and how the market-making institutions are likely to adapt in response.

~

What is important to note about this sort of Marxist “scientific process”, however, is that it looks at broad, abstract patterns first, then looks for ways to explain them—even abstractly—in terms of interactions between abstract objects like bourgeois social institutions and the capitalist class. Although a theory might be too abstract to be very practically useful, it can give us a solid starting point to begin our more empirical, precise research. In this way, we see why Analytical Marxists like John Elster described the Marxist “scientific method” more as a “pre-scientific” method.

In most Marxist empirical research, it makes sense to rely on the normal scientific method of abductive hypothesis formation followed by inductive hypothesis testing (in line with the description of science by C.S. Peirce). But notice what happens in the abductive process—that is, the process of coming up with an initial best-explanation to serve as our hypothesis. In the process of deciding what our initial, best-guess hypothesis will be, the Marxist style of abstract thinking and social analysis has been quite powerful. The Marxist style of materialist dialectical reasoning often manages to provide a surprisingly good first-pass explanation of the observed social phenomenon.

We would, of course, still like to be more confident, by having more concretely worked-out theories developed through empirically-testing. Nevertheless, we often get pretty close by reasoning with characteristically Marxist thinking tools such as institutional selection mechanisms; the corresponding institutional incentives and convergences of behavior; functional explanations; basic Marxist descriptions of bourgeois incentives; and relatively basic observations about economics and human nature which are indirectly relevant to the situation being researched. In lieu of a lot of specifically-relevant empirical research and data, it is impressive to me how accurate these sorts of deductive methods can be.

Furthermore, while predicting a constantly changing, shifting, and dynamic economic system, it can be difficult ever to get enough data to confirm a theory. When empirical testing is essentially impossible, our predictions are forced to rely more heavily on this basic Marxist reasoning style. The best-guess first-pass model is sometimes the best we can ever get, especially if we are trying to decide on actions to take under significantly novel historical conditions.

When rigorous empirical testing or data analysis is not practically attainable, then perhaps the best way of predicting outcomes—possibly through a Marxist style of deductive/abductive reasoning—is as close to a scientific answer as we can get. Arguably, in these special cases, doing this is the scientific method of ‘generating knowledge to our best ability’, and not simply a pre-scientific method of ‘generating hypotheses to test inductively’. However, since we might discover new ways of testing these best-guess statements, I think it is procedurally best to continue considering them as mere hypotheses—but to say that, “Sometimes hypotheses are the best we have to inform our actions.”

~

Our abstract predictions come with abstract caveats and special cases—events like political revolutions, capitalists running into a scarcity of real resources, long-coming crises like climate change—these abstract caveats can interrupt and negate the predictions we make. This is to be expected. Most scientific descriptions of processes come with the caveat that these predictions will occur “all other things being equal”. But the world at the abstract level of social sciences is messy, so the caveats we run into are also messy.

Often, these caveats come from things outside of the immediate model; or occur because our first-pass description of social institutions or classes was too abstract, and failed to capture an internal incentive which turned out to be significant. In these cases, Marxists try to review their old theories to see what they got wrong, and to elaborate their theories to integrate the new nuances and caveats that they had missed previously. Outside spectators of this process often then accuse Marxists of moving the goal posts, and refusing to admit that their theory was wrong. This is sometimes the case, but it usually isn’t—usually, Marxists are engaging in good faith in a very difficult field to be scientifically precise in. In these situations, however, I will admit that it doesn’t help that a lot of Marxists view this dialectical materialist style of reasoning as a special scientific method, rather than a way of making an often-quite-good pre-scientific first-guess.

In practice, when Marxists talk about why their theories did not turn out as predicted, they appeal to caveats made within an even more abstract framework. This can easily come across as if Marxists are arguing for their theory with something very similar to a motte-and-bailey fallacy, whereby a practical issue at a more concrete level gets absorbed into the theory as a mere special case within a vaguer (and thus more acceptable) theoretical framework.

How do we have a rigorous science, which updates our theories, without falling into fallacious arguments? Updating our abstract theories and engaging in a theory-protecting motte-and-bailey-esque fallacy really are procedurally similar. This means that Marxists really do sometimes engage in fallacious reasoning, allowing their politics to cloud their scientific analysis. Marxists should take these sorts of procedural, scientific concerns seriously—as real threats to our practical effectiveness.

To make this discussion of caveats more concrete, here are some examples of caveats to the development of market-making institutions and capitalist monopolies:

• If a crisis is extremely large or drawn out, or other external threats exist, then a government is more likely to take extreme, regime-changing action against capitalist institutions. The government might heavily restrict and constrain capitalists’ freedoms to do what they want in the market—politics constrains capitalism for fascist or socialist ends. Particularly, fascistic ends like nationalism can sustain a market-making institution like the government—at least for a little while—without the government needing to bend their knee to the capitalists. Usually, market-making institutions serve the market players because the capitalists can transform their economic power into political influence. The market-making institutions therefore tend to serve the players in the market by enforcing a predictable set of basic rules and legal guarantees. But if a crisis is too severe, then the market-making institutions might instead get taken over to serve the interests of political opportunists who take power in the disarray caused by a crisis.

• Capitalists are getting smarter about being able to predict how crises might negatively affect them—they worry that too extreme of economic crises can lead to so much social unrest that it hurts their profits. This is likely part of the reasoning behind some in the capitalist class for suggesting a Universal Basic Income, in addition to their concerns about underconsumption caused by people losing basic productive jobs to AI automation. I don’t want to repeat Bernstein’s mistakes, but I think it’s worth noting that this might be a very successful adaptation of capitalism, with the potential to prevent some forms of capitalist crisis for a long time. At the very least, I think this automation+UBI possibility requires a lot more theoretical examination. Perhaps this discussion focuses more on crises caused by underconsumption; rather than those crises caused by capitalists overextending/undercutting our institutions’ capacities to maintain the market, or crises caused by lack of raw resources. (Although perhaps, consumer underconsumption and capitalist overextending are two sides of the same coin. Nov 15, 2021) I might summarize this caveat as my concern that capitalist monopolies might become heavily integrated with their market-maintaining institutions—such that they avoid overextending themselves, and, when necessary, preemptively focus on reliably maintaining and expanding markets—avoiding the sorts of overextending crises described above.

~

Finally, the Marxist style of abstract, dialectical materialist reasoning-to-the-best-explanation can be extended to make more concrete (albeit still best-guess) theories.

Initially, a theory might begin at a high level of abstraction, talking abstractly about the interactions of the capitalist class and the institutions which maintain a capitalism-friendly environment.

But then, we might ask more concrete questions about the capitalist class—what drives them to act the way that they do? We could take an empirical approach here, but this might still be too broad of a question for precise empirical research. Therefore, we can again rely on the Marxist style of reasoning-to-the-best-explanation to create research-guiding hypotheses. We can create an abstract theoretical model about the capitalist class; perhaps describing how its actions as a class in aggregate emerge out of competition and profit-seeking strategies of specific capitalist companies.

By trying to theoretically explain concepts across multiple levels of abstraction, we can create a whole network of theoretical explanations out of what is implicit within relatively general empirical observations. We deduce things using those general empirical observations as premises, and our deductive logical/explanatory tools include reasoning about selection mechanisms, redundant mechanisms, and maybe the occasional functional explanation. These are thinking tools that can help us deduce when situations are likely to converge on particular processes.

When we discover a convergence of reliably consistent material processes, we can use this collection of processes to define an abstract theoretical object. We can use this object as one of our basic building blocks at more abstract levels of our theory. Thusly, we can speak of things like economic classes as actors in the economy, defined by their coordination over class interests. For example: “There is a selection process that, generally, only the capitalist companies which are profitable can survive over time. Otherwise, they would run out of money and dissolve. Therefore, we might say that capitalist companies are selected to care about profit. Thus, established capitalist companies can be characterized as being profit-motivated. (At least, this is the case when they aren’t competing to the death with other companies and trying to undercut each other—which is itself a profit-maximizing strategy in the long run.) To the degree that capitalists coordinate with each other, we can speak of a profit-motivated capitalist class.”

We can thus create a network of theoretical statements: that is, to justify an abstract theory, we can hypothesize other theories at a more concrete level of abstraction. To produce theories at these more concrete levels, we accordingly require more precise empirical data to inform those deductive methods which are characteristic of Marxist abduction. I am reminded of Sellars’ observation on our knowledge of concepts in logical space:

The essential point is that in characterizing an episode or a state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logical space of reasons, of justifying and being able to justify what one says.

— Wilfrid S. Sellars
Empiricism and the Philosophy of Mind, §36

Our knowledge about these abstract concepts in social sciences comes from building the network of theories around them—describing how these concepts interact—in a way which can justify and be justified by other statements in our overarching understanding of the world.

When do we demand deeper (more precise) levels of justification? We tend to want our actions to be able to be justified by our theoretical understanding of the world. This is likely to provoke us to demand more precise justifications of our theories—to match the level of precision that our practical choices demand.

It appears to me that an abstract theory has a difficult time effectively justifying one action over another, when those actions are distinguished from each other at a more concrete and precise level. Abstract concepts usually give conclusions at a similarly abstract level—we might be suggested a range of possible actions, but not know which of those actions is the best to choose. Compelled by the need to pick a satisfactory action from several prima-facie-plausible yet significantly-distinct options, we will hopefully be pushed to develop our theories to greater precision. If we are scientifically motivated, our theories will develop in precision to the degree motivated by the choices we are looking to inform.

Of course, we shouldn’t disregard empirical testing simply because we can use Marxist deductive methods to develop very complex, data-informed theoretical structures. Marxist theoretical statements are developed by discovering patterns which emerge from out of analyzing a lot of data; and which, within our overarching theoretical system, are able to justify and be justified by other theoretical statements. Regardless of whether this integration in our logical space might allow us to construe our theories as a form of ‘knowledge’—our theories will gain much greater precision and practical confidence when we make predictions using those theories, test those predictions, and update our theories accordingly. This sort of scientific hypothesis testing serves to further integrate the reliable theoretical statements and hypothesized processes more deeply into our logical space, while distancing our theories from the less reliable hypothesized statements and processes.

Using the rather powerful deductive strategies characteristic of Marxist theorizing, we can create very elaborate theoretical frameworks about social phenomenon, spanning across multiple levels of abstraction. I’ll have to explore these sorts of deductive strategies in more depth in another post. Even though the deductive method is rather powerful, we still need to engage in real empirical research whenever we can. Empirical, inductive rigor serves as the best test for our theories—it can cut away theoretical ambiguities and vagueness by falsifying some parts of our theories which we once considered potentially valid.

Footnotes & Extra Bits

1 — Luxemburg provides a similar sort of summary of the recent crises of her time, from 1825 to the time of writing. However, she frames it more as proof of the Marxist crisis theory, and less as a ‘charitable reading of history’ as I have done here.

(1) Precise historical analysis of these crises—not simply a layperson’s presentation of history—as well as (2) a theory of common institutional factors and how they consistently converge to form crises; are both necessary to give us more confidence in our theory. However, that may not have been practically in the capacity of left theorists and political organizers back in Luxemburg’s time—for lack of data or for lack of priority compared to overt political action.

Luxemburg’s chapter 2 deserves a lot of praise. Her focus on (1) theoretical analyses of concepts, (2) her analyses’ integrations with other processes/contradictions described by Marxist theory, and (3) her reference to historical events which seem to fit the model—all of these are necessary and useful parts of Marxist deductive/abductive theorizing. However, there may be some issues in how she sometimes uses abstract theoretical patterns to motivate unjustifiably precise predictions.