Brenner’s response to “The Empirical Evidence for the Brenner Hypothesis”

Michael Lind introduced me to Robert Brenner, who has kindly permitted me to publish his response to my post. His response follows.

Anusar Farooqui takes me to task for arguing that the fall in the manufacturing profit rate took place mainly between the mid 1960s and early 1980s, by bringing forward the fundamental point that the buildup of overcapacity in manufacturing did not end then, in the 1980s, but has continued to heighten virtually into the present.   That is of course what I said–viz. that the buildup overcapacity resulting from the continuing intensification of international competition in manufacturing has continued to intensify, with the virtual collapse of investment on a world scale as perhaps its most striking and significant consequence.   But, as Farooqui quickly also concludes, formally contradicting what he just said, the ongoing intensification of international competition and resulting buildup of overcapacity did NOT actually lead to much, if any, further fall in the rate of profit in manufacturing after the early 1980s, which was my contention. Yet, finally, despite the leveling off–not further fall– of the profit rate, there was nonetheless, a deep falloff of capital accumulation, as the overcapacity, intensified competition, and the expectation of the continuation of same profoundly discouraged new investment.

One could say that, in the first instance, the Japanese, then Koreans and Taiwanese, and then their East Asian cohorts above all China, stepped up capital accumulation far more than was justified by the immediate/static profit rate, because they knew that, given that US and other earlier developing producers could not much reduce costs because production technology was frozen in the existing capital stock and wages were pretty inflexible, they could sell more cheaply than the Americans on the world market for as far as the eye could see (in practice, until the huge devaluation of the dollar from 1971-1973-1979).   On the other hand, world producers have, over time, reduced investment growth to an ever greater extent, because they have known that, at whatever costs they are able to produce at, will soon be bettered by lower cost producers coming on line.   Capital accumulation was amplified because the Japanese especially but other latecoming investors knew that the earlier comers could not match the latecomers costs.   Capital accumulation slowed down because all manufacturing producers knew that later developing entrants would be able to throw commodities onto the world market that were cheaper than they could possible match.   So after the Great Recession of 2007-2009, even the Chinese, overwhelmed by over-capacity and the collapse of world demand growth, could not prevent investment in China from plunging, because the firms there knew they could not match the costs of even later entrants (Vietnamese, etc) and could never make up for the drop off of US borrowing and US debt-driven purchases of Chinese goods.

Farooqui goes on to tax me for seemingly failing to register that, while the manufacturing rate of profit fell sharply, manufacturing makes up only 12% of value added, and the overall rate of profit did not fall.   To show this, he adduces a graph that shows a major recovery, if still incomplete recovery, of the corporate rate of profit.     But I think he is wrong about this, because I don’t think that the corporate rate of profit is a good measure of economy-wide profitability, since it includes the profit rate of the corporate financial sector.   The financial sector rate of profit can’t be properly included in the measure of the corporate rate of profit, because the financial rate of profit is not properly measured, as it is in the BEA data (and Farooqui), as financial profits over financial capital stock; it should rather be measured as returns from investment in financial assets (mostly on paper).   So, the corporate financial rate of profit must have financial assets overall, not the corporate financial capital stock, in the denominator, just as it must have financial returns in the numerator.

So, to properly get an idea of profitability or the rate of return for the economy as a whole, one must confine oneself to the profit rate for the non-financial corporate sector, which is properly measured by non-financial corporate profits/non-financial capital stock. If one consults this latter measure one can see that, properly measured, the recovery of the profit rate for the economy as a whole is still pretty far from complete.   What has happened is that the fall and failure of the manufacturing rate of profit has brought down the overall rate of profit, but not nearly as much as in manufacturing itself, because there was no significant fall in the rate of profit outside of manufacturing.   It is the divergence between the trajectories of the manufacturing rate of profit (sharp fall) and the non-manufacturing rate of profit (basically flat/stable) that was my main evidence for arguing that the fall in the rate of profit were the result of what happened in manufacturing almost by itself…and thus an intensification of international competition that obtained only in manufacturing but not in nonmanufacturing. Bottom line is that the economy wide rate of profit did fall, if not nearly as much as did the manufacturing rate of profit.

Farooqui, in bringing his article to conclusion, puts a big emphasis on the big shift in the wage share against labor and in favor of capital, which began in the 1970s (if not a bit earlier).   Here I agree with him completely, even if I think that this rise in the profit share does not fully restore the rate of profit for the economy as a whole.   Indeed, I would take this point much much further, arguing that the gains for capitalists (and the rich) are nowhere near fully accounted for by what might be called the effect of austerity on wages (and government services), but are, especially from around 1980, to be found above all in the direct political interventions of both political parties in alliances with top corporate managers (especially in finance) to radically shift– by way of taxation, returns to lending to government, etc etc– income and wealth upward to the one per cent or, really, top 0.1% or top 0.01%   At least that’s what I tried to argue in the section on neoliberalism in my editorial in Catalyst, number 1.






What explains the tactical superiority of the Wehrmacht?


The tactical superiority of the Wehrmacht is evident from the fact that while the loss exchange ratios became more favorable to the Soviets during to the course of the Soviet-German War, the Soviets did not, in fact, manage to close the gap. Why might that be? Glantz and House have shown convincingly that by 1944 the Red Army had mastered the operational art of combined arms mechanized warfare. Red Army operations acquired a tempo, depth and confidence that Guderian surely envied. Why then did the exchange ratios not become favorable to the Soviets in the third period of the war?

Traditional explanations of German tactical superiority over Allied armies appeal to the quality of German officers and to institutional features of the German army. Commanders in the field were issued missions to be accomplished instead of detailed instructions. The exact means of accomplishing the missions were left to the field officer. This encouraged initiative and was more efficient than micromanagement of army operations from faraway headquarters. The quality of the German officer corps was thought to be superior, both on account of Prussian military culture as well as the special experience of the Versailles era. The latter idea being that since the Reichswehr was limited to 100,000 personnel and hence useless for actual defense, it essentially became a training institution for future officers. I am not entirely unsympathetic to these explanations. But I think there is a more straightforward one.

The main light machine gun of the Wehrmacht was the MG-34, later replaced by the MG-42. The latter was so successful that when the Bundeswehr was founded in 1955 and issued American firearms, experienced German soldiers found them wanting. But all existing designs had been destroyed in 1945, so Rheinmetall was forced to reverse-engineer the MG-42 from existing specimens. Modern variants of the MG-42 continues to be used by the Bundeswehr. And the machine gun continues to see action in places like Syria. The history of this weapon is so interesting that Adam Tooze wrote a paper on it.

The MG-42 was a 7.92mm caliber gun. It was 1.2 meters long and weighed 11.5 kilos with the bipod; which meant that it could be carried on the shoulder. It had a high muzzle velocity of 755 meters per second. But what made the MG-42 a supremely effective machine gun were three features. First, it had a very high rate of fire; about 1,200 rounds per minute. Second, it had an effective range of 2,000 meters with a bipod, and 3,000 tripod. And with a tripod, it could even be used to deliver indirect fire. Third, and perhaps most importantly, it was belt-fed. Moreover, the belts could be linked to maintain continuous suppressive fire. This made the MG-42 (and the MG-34 before it) an excellent universal machine gun. That is, it was light enough to be carried by one man; thus serving as a light machine gun. But when used with a tripod or mounted on tanks or combat aircraft, it became a heavy machine gun that could project continuous suppressive fire at a great distance. With a special mount, it even doubled up as an effective anti-aircraft weapon. And above all, it could sustain suppressive fire indefinitely; constrained only by ammunition supply.


MG 42

The Soviets had combat aircraft, tanks, and artillery guns that were comparable or even superior to the Germans. They even had light machine guns that could be carried but they were fed by 71-round drums (the Finns called them ‘record players’). They also had heavy machine guns. But none could be carried into battle. The big ones had to be pulled by horses; the smaller ones had to be dragged on wheels. The Soviets did not make a single belt-fed machine gun that could be carried into battle throughout the Soviet-German War. It was only in 1946 that the Soviets finally introduced their first such machine, the RP-46.


The Red Army’s SG-43.

The German Army built the MG 34 and MG 42 into the core of its infantry organization, right down to the Gruppe (squad) level. The most basic fighting unit was the rifle squad. The rifle squad’s primary firepower did not come from rifles; it came from the machine gun. The job of the squad leader was to direct the emplacement and fire of the machine gun. The machine-gunner (No 1 Man) was responsible for carrying, firing and maintaining the light machine gun. He also had to carry a 50-round belt drum. The assistant machine-gunner (No 2 Man) carried a spare machine gun barrel, four 50-round belt drums, and one 300-round ammunition can. Another spare barrel would be carried by the ammunition carrier (No 3 Man), together with two 300-round ammunition cans. The other six riflemen often carried additional ammunition for the machine gun. In total, the squad could carry 5,000 rounds on their persons. The MG-42’s reload mechanism made it run faster and faster as you sustained fire; until it approached about 1,500 rounds per min. To be sustained at this rate for any appreciable length of time required prepared positions with stocks of ammunition; whence the dramatic tactical advantage of strong points at the ends of effective supply lines; and, of course, back to the great machine working behind it to reproduce the whole show.


Adam Tooze notes that the job of the German rifle squad was to work the machine. It was the reason for their existence as a unit. The survival of the riflemen depended on their working together to feed the machine whose fire protected them on the battlefield. This is what accounts for their cohesion. An answer to why did the Wehrmacht have to annihilated? Why did it hold together? Did National Socialist Germany hold together until the Battle of Berlin because it was a machine civilization?

My question is low-stakes. What explains the exchange ratios in the Soviet-German War? At the basic tactical level, whereas the Germans ran their factories of firepower, the Soviets did not even have an infantry squad (their basic tactical unit was too large; the size of a German company) for the obvious reason that they did not have a real machine gun. But Stalinist USSR was also a machine civilization.

German superiority at the tactical level (roughly, squad-company) was defeated not only by the Soviets’ superior performance as a machine civilization, but also by the fact that the Red Army became better than the Wehrmacht at the operational level that came to dominated by combined arms mechanized warfare. The Red Army rediscovered ‘deep operations’ in early 1942. Glantz and House show how the Soviets and the Germans exchanged places in mechanized warfare during 1942-1943.

During the Battle of Stalingrad and the ensuing winter campaign, the tank and mechanized corps organized in 1942 had proven their worth as instruments for the limited tactical exploitation of enemy rear areas. For the remainder of the war, high-priority combined-arms armies such as the guards armies would control one or two tank or mechanized corps for the purpose of encircling German defenders to a depth of 50 to 200 kilometers behind the front lines.
         However, the Red Army needed a larger mechanized formation, analogous to a panzer corps or panzer army, of deeper operational exploitations up to 500 kilometers. The result was the 1943 tank army.…The depths to which these armies penetrated increased steadily throughout the war.…For the remainder of the war, the five (later six) tank armies were the spearhead of Soviet deep attacks, conducting operational maneuver and seeking objectives deep in the German rear areas. On a map, Soviet offensive plans often resembled a set of nesting dolls, with shallow encirclements inside of other deeper encirclements. The separate tank and mechanized corps, sometimes replaced by cavalry-mechanized groups in difficult terrain, were attached to the forward combined-arms armies so that they could encircle one or more German corps immediately behind the German main defense lines. Meanwhile, operating under front control [a front is the equivalent of a German army group], the tank armies bypassed these struggles, straining to penetrate as far as possible into the operational depths and thereby achieve larger encirclements.

Glantz and House (When Titans Clashed, p. 207)

The persistence of the Germans’ tactical superiority even as Soviet ‘tank armies’ gained operational superiority is interesting from the point of view of struggling machine civilizations. Perhaps even in purely tactical terms, the Germans were saved by the bell as it were for the RP-46 production came online in 1946. But in the lived experience of Soviet-German war, only one side had the machine gun. And that is an efficient explanation of German tactical superiority.

P.S. Why, of course, the material culture of the Wehrmacht is fascinating! Look at these men covered with food for their hungry machine.


A work gang covered in raw material.


Thinking through connections between inequality, home prices, schooling, and taxation.

(Thanks to the Policy Tensor for welcoming this guest post.)


I have a strong intimation that one of the major drivers of inequality in American society relates to land use. Here is how it works.

Let’s start with labor markets.

We live in a world characterized by extremely high returns to certain forms of educational achievement. (Whether that achievement has to do with ability, or with social/cultural capital credentialing, I do not venture to say.) Individuals who have passed through certain schools and acquired certain degrees have open to them employment in extremely high paying jobs. Indeed not only are the incomes of those at the top high, they are also increasing at a rapid rate, creating an ever-growing gap between those in the highest paying jobs and those in nearly all other jobs in the economy, where wages have stagnated, if not declined.[1]

The increasing returns to those credentials, in a world where their supply is effectively fixed, naturally increases competition for access to those credentials.

Beyond the very top earners, throughout the economy there is a widely documented increase in the premium to education – that is, the wage differential between college graduates and those without a college degree is large and growing. (Autor, 2014) Though I have not seen direct evidence documenting it, this premium must increase markedly with the imputed quality of the college education received, almost certainly proxied by “brand.” Combined with the fact that since 2000 the only sector of the economy in which total employment has increased is the low-wage sector (Autor, 2015; Mishel and Bivens, 2017), it’s easy to account for a sense of scarcity among parents over the supply of future remunerative employment for their children.

In theory it is high school aged kids who are competing for these scarce college placements, and in turn for scarce jobs. But in practice children have very little say over how they perform in this competition. Instead the key factor is their parents.

If parents detect that there are increasing returns to certain forms of credentials, and perhaps correspondingly fear that failure to gain access to those credentials may doom their children to low wages (setting aside for the moment status competition between the parents and their peers, which may account for a great deal), then a rational parent wants to undertake whatever investments in their children’s education will best position them for access to these highly coveted spots.

What are the key investments that parents can make?

Some of the most valuable investments are likely to have been made a long time ago – for example, the parents’ own education, as well as the parents’ health, incomes, etc. But from the time the children are born, one of the most important factors that parents can control is where they send their kids to school. High quality schools correlate with better college placements. (At least I think so – I have not seen direct evidence of this but spend 10 minutes with any parent today and you can be sure that parents think this correlation exists.) Most American children go to public schools, and most of those public schools are in the same neighborhood or school district as the household’s residence (75% in 2007, at least, according to US Department of Education, 2010). Therefore the mechanism by which most parents invest in their children’s schooling is by choosing where to spend the portion of their income that they devote to housing services. Since most American households own rather than rent their homes, this is predominantly a question of where to buy a home. It stands to reason that home prices should be higher in places that have good schools.

Indeed, studies show strong correlations between school quality (measured for example in terms of average test scores, which happens to be the data generally most easily accessed by parents comparing schools) and a premium paid on home prices (e.g. Chiodo et al., 2010; Nguyen-Hoang and Yinger, 2011). Since Tiebout’s famous paper in 1956, the principle way to model home prices has been as a function of a distribution of demand for amenities, including government services. Those amenities that are most valued translate into the highest home prices. Schools are clearly one of them.

The sociologists Reardon and Bischoff (2011) have done some very interesting work showing that as income inequality has increased in the US, so has the residential segregation of affluence. Since income inequality in the US has a pronounced “upper-tail” quality – that is, it has rapid growth in earnings at the very high end – the small numbers of very high earners increasingly bid to live in a small number of highly desirable locales, one aspect of which is the presence of very good schools.

In a 2016 article, the education sociologist Ann Owens makes the striking finding that residential segregation by income in America is increasing – but that the effect is overwhelmingly limited to households with children, which make up only 1/3 of American households. (Owens, 2016) Households without children seem relatively content to stay where they are. Families with children move, and they move in ways that segregate them. In combination with the results of Owens, Reardon, and Jenks (2016), which shows growing income segregation between schools and school districts, it seems clear (and, again, is intuitive to anyone who has spent time with the parents of young children) that parents who intend to enroll their children in the rat race of educational achievement make very careful decisions about where to live so as to ensure that their children go to good schools. They pay a premium to be able to live in such places. The more income inequality there is, the higher the premium to be able to live in the places with the best schools, the more segregated those communities and their schools become, and – and here is of course the kicker – the more that those already successful are able to monopolize access to those public goods which are likely to make their children successful.

So if you are someone who thinks that this is a perverse state of affairs, what types of remedies might be available?

1) Big picture you would want to do something about the organization of the American economy that produces such lopsided, pre-tax and transfer income distribution in the first place. But let’s set that aside for now.

2) If you could not change (1) you would at least want to have a radically progressive tax system such that even a skewed pre-tax and transfer income distribution became less skewed post-tax. If you taxed away much of high earners’ incomes, it would presumably limit the extent to which they could bid up house prices, and this might in turn limit their ability to exclude those who earn less. But let’s even set this aside for now.

3) One thing you might do is build affordable housing (or create large numbers of housing vouchers) for poor and working-class families to be able to move into affluent school districts. This would be a good thing, but as the land in question is already expensive, it would be a hugely costly proposition. If enough housing was built (all of which, incidentally, would have to be built over the screaming protests of the existing residents) it might turn out to be counterproductive, in that some portion of the affluent would choose to flee to further out suburbs, and by taking their property tax payments with them, they could contribute to a worsening of the schools they left behind. Still building affordable housing in affluent communities is almost surely a key ingredient in any overall recipe for reform.

4) Neoliberals (e.g. Glaeser, 2017) believe that one of the most important reforms you can make is to reduce land use restrictions that prevent the high demand to live in certain communities from resulting in the construction of more housing. So do away with local ordinances that require lots to be a certain size or that prohibit the construction of multifamily housing. Something like this could work in tandem with (3).

5) One thing you could definitely do is reform the tax code to reduce all of the provisions that incentivize taking on massive leverage to buy a home. The most obvious is the mortgage interest deduction, which if eliminated would substantially lower house prices. (Harris, 2013) (The Republican tax plan lowers the ceiling from $1,000,000 to $500,000, but only for new purchases. That probably has the effect of raising home prices by discouraging anyone who is currently taking advantage of a large deduction from moving, thus limiting the supply of houses for sale.) The best idea I’ve seen would be replacing the mortgage interest deduction with some sort of equity subsidy (as suggested by Stiglitz, 2014), especially if, as Stiglitz suggests, you just establish a blanket $100,000 income tax deduction for everyone. That would ensure that working and middle class homebuyers don’t lose out. Another thing it would be good to eliminate (again, as long as you have something like a $100,000 flat deduction) is the property tax deduction. If you were not able to write off part of the cost of paying for your children’s education, it should reduce the amount you are willing to pay for a house, and that should also slow home price appreciation.

6) What I am really trying to figure out is if there aren’t more draconian measures one could take to break the connection between where you live and the schools your children go to. There are of course (or, there have been) busing programs of various sorts. In the era of court-ordered desegregation, however, these programs tended to be limited to within-district integration. The result was to contribute to white flight, since parents who could afford to preferred to move out of cities rather than keep their kids in integrated schools. Today, there is the “Moving to Opportunity” program, which gives vouchers to low income families from high poverty neighborhoods to move to more affluent areas and go to school there. (Chetty et al., 2016) This is good for those kids, but does very little to limit the ability of affluent parents to segregate themselves for the purpose of attending segregated schools, which is their main objective.

What if you just ended local control of schools and instead created large, integrated regional districts, large enough that the distance affluent parents would have to move to get outside them created a disamenity so great that it overwhelmed the gains to accessing segregated schools? One can imagine a general equilibrium scenario in which the response to this sort of a change is employers moving jobs really far outside of metropolitan areas in response to the demand of high-skill workers to live in segregated communities. But employers want to attract workers without kids as well as workers with them, and workers without kids want to live in the heart of thriving cities, not in ranch houses 100 miles outside them. So there would be a lot of forces working against an outcome like that. In the meantime, genuinely integrated regional school districts could create a degree of randomization in school assignment. In a “veil of ignorance” type situation, affluent households should support a tax regime that ensures that all schools have adequate funding.

So, those are my proposals for today: build affordable housing in affluent communities; reform the tax code to eliminate those features which incentivize bidding up home prices; regionalize metropolitan schools and randomize school assignment within them.

What am I getting wrong?

[1] Interestingly, the perception itself may be anachronistic. In their recent work on distributional national accounts, Piketty, Saez and Zucman (2016) find that between the 1970s and the 1990s the growth in the incomes of the top 1% of earners was driven by increases in labor income (e.g. salaries). Since 2000 it has been driven by growth in capital income (e.g. primarily earnings on stocks, bonds, and real estate). Some of this may be misleading – for example many high earners are paid in part in stock options and the like, which look like capital, but are in effect labor compensation. Either way it probably does not make much of a difference from the point of view of parents investing in human capital. There are still only a small number of jobs that will pay your children big enough salaries to be able to accumulate the capital assets from which they will then derive capital income.


Does Mohammad bin Salman want to be a Sultan or a Despot?


After the oil price revolution what emerged in the gulf was a distinct model of a ‘rentier state’ that Gause called the ‘Oil Monarchy’. The oil windfall meant that the state came to enjoy a radical autonomy from society. With more than half the population on the state payroll, 100 percent of Saudi citizens dependent on ‘cradle-to-grave’ welfare, and almost all Saudi businesses feeding directly or indirectly at the state’s tits, the Al Saud came to exercise uncontested control over the country. The Wahhabi clerics seemed to enjoy a degree of autonomy but in reality they too were tamed. Power in Saudi Arabia rested firmly in the corporate body of the ruing family, with different factions or subbranches competing for influence in what was effectively a ‘ruling oligarchy’ à la Winters. Inside the corporate body of the family, the King was no more than a ‘first among equals’. American protection after the Islamic Revolution provided external security, letting the Al Saud concentrate on family politics, propagating salafism worldwide, patronizing clients like Pakistan, and playing the regional game.

Crown Prince Mohammad bin Salman has moved decisively to consolidate all power in his own hands. There are two theoretical models available to understand these developments. In Winters’ frame, what we have is a shift from a ‘ruling oligarchy’ to a ‘sultanistic oligarchy’. Such shifts are attended by purges and highly public disciplinary attacks by the dominant oligarch against his rivals. When a dominant oligarch manages to establish his supremacy, he can arbitrate disputes amongst the oligarchs, thereby generating consent and stabilizing his rule. If he fails, the bid is likely to be generate instability. We have seen this process unfold more or less transparently in Russia under Putin, India under Modi, and China under Xi. It remains to be seen whether Salman can pull it off in Saudi Arabia.

The other theoretical model I have in mind is developed by Kotkin in Waiting for Hitler, the 1200-page second volume of his biography of Stalin. In that frame, the shift is from dictatorship to despotism. A dictatorship is established when a clique gains unchallenged control in an autocracy; instead of power being dispersed in multiple institutions as in legal-authoritarian orders such as Singapore, power in a dictatorship is unchecked and concentrated in an informal network at the very top of the hierarchy. In Stalinist USSR, this clique was a ruling group of fewer than a dozen individuals who usually met at Stalin’s dacha. They constituted a veritable ‘state within the state’ and made all the important decisions of the party-state in the name of the Politburo. The shift from dictatorship to despotism occurs when the dictator effectively neutralizes all his rivals in the ruling group. No more ‘first among equals’, the despot consolidates his power at the expense of the rest of the ruling group. Stalin moved towards despotism after the assasination of his closest friend Kirov. In order to do so, he had to unleash the terror.

Although both lenses are illuminating, I find Kotkin’s frame more useful. For it allows us to go beyond the oligarchy and view the shift in a larger frame. Specifically, it allows us to also triangulate the mass politics of dictatorship. Mass politics means that the dictator appeals directly to ‘the people’ for support, bypassing mediating interests. Often this takes the form of populism and a ‘cult of the leader’. Moving from dictatorship to despotism exposes the dictator to conspiracy by his threatened rivals—it must therefore be seen as a sign of weakness; whose origins lie in the dictator’s real or imagined insecurity. A resort to populism increases the dictator’s personal power relative to his rivals in the ruling clique. The interaction of the bottom-up politics of resentment of the elite with the dictator’s bid to achieve despotism can generate tremendous violence and instability. Herein lie the origins of the terror.

At this early stage it is not in fact clear whether Salman wants to a sultan or a despot. Perhaps he himself doesn’t have a specific end-state in mind. But the will to power is strong in this one so it is inadvisable to assume that he has moderate ambitions. What we can say is that, as we watch the shift unfold in Saudi Arabia, we need to pay attention not just to the Crown Prince’s consolidation of power within the ruling family and over the Saudi rentier state, but also for signs of mass politics that attend a shift from dictatorship to despotism.


What to Do in the Face of Mass Technological Unemployment

The risk of job loss is the principal risk faced by the subject of a liberal market democracy. In the era of ‘social democracy’ 1946-1979, countercyclical fiscal policy was combined with unemployment insurance to counter the risk; in the ‘neoliberal era’, that is to say since Volcker’s appointment, the Fed has taken the lead in macroeconomic stabilization. Both of these solutions were hit upon to deal with cyclical threats; the first to deal with high unemployment; and the second high inflation. Neither is a solution to a secular rise in the unemployment rate. While more generous unemployment insurance and better macroeconomic stabilization may help in a modest uptick in unemployment due to labor saving innovations, they would be insufficient to the task were mass technological unemployment to arrive in force. By mass technological unemployment we mean the Luddite nightmare—a real shortage of work for a painfully substantial portion of the populace due to labor saving innovations.

Say’s Law is usually deployed to point out the so-called Luddite fallacy. But Say’s Law only holds only if one assumes that the real wage will never fall below the subsistence level; thus allowing the labor market to clear without unleashing mass starvation. Even with a completely flexible price mechanism, there is no logical reason why the market clearing wage rate could not fall below the subsistence level. Hence there is no logical reason why the Luddite nightmare cannot obtain.

Yes, historically that did not happen. Work did not disappear as a result of labor saving innovations, including the introduction of radically new machines and techniques. It may be that some yet-to-be-discovered deep law of economics forbids it. But it could also be that the rapidity of technological change has nonlinear effects. And that the revolutions associated with steam, telegraphs, railways, internal combustion engines, telephones, computers, and the Internet, were just not revolutionary enough. That a major technological revolution—such as a rapid near-total automation of all existing work—may yet give life to the Luddite nightmare.

In what follows we will assume that it has arrived. The question is: What should we then do about it? This is posed in the technocratic spirit of Bernanke (2004)’s proposals on what to do were the economy to hit the zero lower-bound for a prolonged period. His proposals became the basis of the highly successful policy innovations introduced by him; later sustained under Yellen; and copied by all hard currency central banks.

Unfortunately, Bernanke’s technocratic solutions which rely on elevating the wealth of investors in search of ‘wealth effects’, are unlikely to work in the event of the Luddite nightmare. The monetary authority may find itself fighting deflation and high unemployment at the zero lower-bound; pushing on a string for years on end. A major technological revolution could make monetary management of the economy completely unviable, thus ending the primacy of the Fed. Governments would quickly find themselves too indebted to keep fighting unemployment rates still hovering above 30 percent. Markets would lash out and punish sovereigns, now paying no more attention to the impotent central bankers. People would not accept being abandoned to market discipline. The whiplash of the world economy is likely to generate significant political instability both in the center and on the periphery.

The good news is that crises can be useful for pushing through game-changing policy innovations. My wager is that with a combination of legal, institutional and policy changes it is possible to shield the populace from the whiplash of the global market without dismantling the open economy and the liberal order.

My proposal has five components. All five are important and they will work best in tandem with each other.

First, the legal and effective workweek will be reduced from five eight-hour-days to four; with mandatory 30min lunch breaks and employers required to pay overtime for any and all work done over 30 hours in any given week. The impact of this measure would directly redistribute work from those who have to those who don’t by a simple intervention with minimal distortions. Apart from generating a substantial near-term reduction in the unemployment rate, the utility gain in leisure time for the entire population would be substantial. It might reduce labor productivity by letting second-benchers do as much work as the first-benchers, but that should be the least our worries with labor productivity so high as to precipitate mass technological unemployment.

Second, and to compensate for the first, the minimum wage rate will be raised to be consistent with generating enough money to provide a decent standard of living for a family of four with two working adults now working only 30 hours a week. If you think $15 is fair for the US at the present time, then under my proposal, the minimum would have to be $20 an hour. (This can be implemented fairly easily through market mechanisms if my next proposal is adopted.)

Third, the unemployment dole will be abolished tout court and so will unemployment itself. A National Labor Service (NLS) will be formed. It will pay the minimum wage in exchange for work to every legal resident who so desires. Each person who signs up will list their skills and provide their resumes—and they will be given priority for jobs that match their skillset. Labor allocation will be tasked to an independent agency within the NLS called the Planning Board. It will coordinate the allocation on ‘the supply side’ with the Works Department (‘the demand side’). The Works Department will be given priority as a subcontractor at the federal, state and municipal governments. The surplus labor pool, those who are not absorbed by the Works Department, will perform community service, receive skill-augmenting instruction and training—including military, first aid, and emergency response training. The point is to harness the labor surplus to augment national strength. In order to make the NLS an effective instrument of the national will, Congress will exercise oversight and authorize funding. (My last proposal addresses the question of public finance.)

Fourth, once unwanted unemployment is effectively abolished by the NLS, monetary and fiscal authorities will be freed from their mandate to maximize employment. The Fed’s mandate will change from ‘maximize employment subject to monetary-financial stability’ to ‘maximize real wage growth subject to monetary-financial stability’. The fiscal authority, ie Congress, will be freed entirely from macroeconomic stabilization. It will instead focus on negotiating traditional questions of redistribution (taxation and tax expenditures) and the new game of fine-tuning the NLS.

Fifth, in order to pay for the NLS, Congress will institute a wealth tax in line with Piketty’s proposal. For if robots do all the work, that’s where all the surplus of the society will be concentrated. (To find the money, you go where it lives.) In principle, one could increase corporate income or capital gains taxes instead. But Piketty’s tax proposal is less distortionary (since, unlike income tax, it doesn’t affect any agent’s allocative choice at the margin) and better targeted (a wealth tax is more progressive than income tax because wealth is an order of magnitude more concentrated than income). Congress will also need to abolish all unnecessary tax expenditures. These giveaways come to $1.5 trillion in the current fiscal and some 70 percent of them go to the top 40 percent. Closing these loopholes will free-up a trillion dollars of revenue. This will augment Congress’ ability to pay for the NLS without running up public debt.

In a work scarce world, it’ll be desirable to spread the work (and the leisure) around. My 4-day workweek proposal addresses that problem head-on in a straightforward and simple-to-implement fashion. Raising the minimum wage is necessary for the reduced workweek proposal to work. The NLS is my core proposal. In a truly work scarce world, some form of public works program will become indispensable. It would be best if we have a rational plan for it in advance. The pooling of risk creates value through diversification. The gain is systemically emergent in the sense that the source of value cannot be located in any of the units—the act of pooling of the risks faced by the units itself creates value. With the NLS as an instrument of national will, this gain would accrue to the nation-state. It will also free-up policy room at the Fed, which I suggest should then be issued a fresh mandate to maximize real wage growth subject to monetary-financial stability. Finally, the NLS will have to be financed from the public purse. I support Piketty’s tax because it is more efficient than the alternatives.

A bold combination of reduced workweek, work guarantee, minimum wage, and wealth tax can work even in the teeth of robotic competition in the labor market. Mass technological unemployment poses a significant threat to political stability. The solution to the problem of excess labor is to absorb it. Some form of work or income guarantee will be necessary for social survival in the nightmare scenario. The advantage over guaranteed income is that you are not paying for sloth. Is a modern civilization without work even possible? Shouldn’t we ask for something in the public interest in return for public money? And shouldn’t surplus labor be transformed from being a source of weakness and imbalance into a national asset? An instrument of national will? Surely the United States could not be hurt by training millions of men of military age in reserve battalions rather than incarcerating them?


War Production in the Nazi-Soviet War

The Second World War is an ideational construct of the Western imaginary; one that bundles together a titanic great power struggle on the continent with a bunch of sideshows in the Atlantic, Mediterranean and the Western Pacific basins. The main turning point in the history of the twentieth century was not World War II but the Soviet-German war. It was above all the Red Army that fought, defeated and annihilated the Wehrmacht. It was the outcome of this struggle that decided who would join the Anglo-Saxon powers at the top table.

On the eve of Operation Barbarossa Western intelligence expected the USSR to knuckle over under the German onslaught. The Soviet Union was thought of as ‘modernizing’ but not modern; Soviet industry was seen as deficient in quality and productivity; the Soviet economy as grossly inefficient and disorderly; the Soviet workforce as lacking in skill and knowhow; Soviet administrative apparatus as shot through with red tape; and Soviet transport as characterized by severe bottlenecks. The Red Army was seen inferior to first class militaries and devastated by the great purges; perhaps good for defense but certainly not attack. And most importantly, there were definite questions about the stability of Stalin’s regime; whether it enjoyed support among the masses; or whether morale would collapse and take the regime down with it in the event of a Nazi invasion. In short, Soviet modernity continued to be seen as partial and incomplete in Western eyes.

Even after the Red Army parried Operation Barbarossa, Western intelligence agencies continued to worry about Soviet collapse or a separate Nazi-Soviet peace; especially given that perhaps more than a third of Soviet war potential was located in the in the Soviet territory now occupied by Germany. Indeed, it was not until the summer of 1943 that the Soviet Division of ‘the O.S.S. chairborne’ finally became confident of a Soviet victory and recognized the Red Army as the strongest land army in the world. The Division then began advocating for a ‘second front’ and for a modus vivendi between ‘the Big Three’ for the postwar world that would remain its idée fixe until 1945. However, it was already clear by the winter that Operation Barbarossa had failed and that the communist great power would ultimately prevail. Why then did the Division wait until the summer of 1943 to ‘call the war’?

We will come back to the question of Western perceptions of Soviet strength in a future essay. In what follows we will instead examine the balance of power in the early-1940s and offer some observations on the sources of Soviet strength.


Figure 1. The revealed balance of power in the early 1940s.

Figure 1 shows the armament production of Nazi Germany and the Soviet Union as a percentage of US production. We know that Soviet war production was constrained by coal production and transport bottlenecks rather than plant capacity; which was the case in the United States. Yet, the average Soviet:US ratio in armament output was 1:1.4. So the US was decidedly stronger than the USSR but not by all that much. By comparison, the average German:Soviet ratio was 1:1.5. So the Soviet-German gap was roughly the same as the US-Soviet gap.

Put another way, this German:Soviet:US ‘balance of armament power’ ratio is 1:1.5:2.2; which is much more asymmetric than the Correlates of War Index ratio of 1:1.3:1.2. Our narrower balance of armament power ratio better captures the actual balance of power. For while one may quibble with the exact ratios, there is no denying the quantum power gaps between the US, the USSR, and Germany in 1941.

During the three full years of combat in the Soviet-German War, 1942-1944, the unoccupied Soviet Union produced 78,000 tanks and self-propelled field guns and 85,000 combat aircraft; National Socialist Germany produced 35,000 and 65,000 of these two war machines respectively. Table 1 reveals that despite the loss of substantial arable land, industrial base and skilled population to German occupation in 1941, the communist great power comfortably out produced Germany not only in tanks and warplanes but also in rifles, machine guns, and artillery.

Table 1: War production, 1942-1944.

Rifles Machine guns Guns Tanks Combat aircraft
Soviet Union (‘000) 9,935 1,254 380 78 85
Nazi Germany (‘000) 6,501 889 262 35 65
Soviet/German (ratio) 1.5 1.4 1.5 2.2 1.3

How was the Soviet Union in a position to produce such prodigious quantities of armament? Stalin had ordered the construction of massive excess capacity in the armament industry. But given the binding constraint of scarce skilled labor, capacity wasn’t enough. What was required was a productivity miracle.

By 1944, Soviet defense output nearly quadrupled to 3.9 times the 1940 level. But employment in the defense industry grew by a factor of only 1.6 (from 1.8 million to 2.9 million). The difference was accounted for by a veritable productivity miracle in the armament industry. In 1940-1944, output per worker in the Soviet armament industry trebled as mass production of field-tested armaments (the T-34 and the Sturmovik above all) was scaled up. Speer’s ‘armament miracle’ is much less impressive by comparison. By 1944, German output per worker grew to only 1.8 times its level in 1940.


The Soviet margin of superiority documented in Table 1—with a median of 1.5:1 across modern weapons—would not have obtained without the productivity miracle of the 1940s. Table 2 marshals the evidence for this claim. We see that had productivity stalled completely in Soviet armament production—as it did in agriculture and civilian industry; see Figure 2 above—Germany would’ve enjoyed an overwhelming preponderance in all armaments; with a median ratio of 1:1.9. Even if Soviet productivity had grown at the German pace, the ratios would’ve been squarely favorable to the Germans; with a median ratio of 1:1.3; a level sufficient to reverse the odds in the Soviet-German War.

Table 2: Armament ratios under different Soviet productivity growth hypotheses.

Soviet:German Rifles Machine guns Guns Tanks Combat aircraft
Actual 1.5:1 1.4:1 1.5:1 2.2:1 1.3:1
Soviet war production assumed to grow at the German rate 1:1.2 1:1.3 1:1.3 1.2:1 1:1.4
Assuming zero productivity growth in the Soviet Union 1:1.8 1:1.9 1:1.9 1:1.2 1:2.1

Figure 2 displays the Soviet margin of superiority under the three productivity hypotheses. Again, it is clear that a productivity miracle was necessary for the Soviet margin of superiority to hold.


Figure 2. Margin of Soviet superiority under different productivity hypotheses.

Our hypotheses on Soviet productivity are uniform across weapons. If we drill down to the level of major weapons, the importance of Soviet productivity growth becomes even more evident.

Midcentury mobile air-land warfare singled out the medium battle tank and the dive bomber as the principal weapons of war. Once good design solutions were tested on the battlefield, they had to be mass produced in large numbers to take advantage of economies of scale. The Germans settled on the Panzer III and IV for their Panzer armies and the Stuka for their dive bomber. The Soviets settled early on the T-34 as their main battle tank and the Il-2 Sturmovik as their dive bomber of choice. Both the T-34 and the Il-2 were superior to their German counterparts.

Table 3. Real unit cost of producing principal Soviet war machines.
Source: Harrison (1996)

Il-2 Sturmovik T-34 medium tank KV heavy tank
1941 330 270 635
1943 162 135 225
1943/1941 49% 50% 35%
Average 45%

Table 3 shows that the real cost of producing the Sturmovik and the T-34 fell by half in just two years. The costs of producing the KV heavy tanks—required for infantry support and urban fighting—fell by two-thirds. The real constraint on production was not cost but scarce labor. Table 4 displays the average number of labor hours required to make each unit in 1941 and 1943.

Table 4. Direct labor requirements of war production.
Source: Harrison (1996)
1941 productivity level 1943 productivity level
Output (thousands) Hours per unit Total hours (millions) Hours per unit Total hours (millions)
Rifles 9,935 12 119 9 89
Machine guns 1,254 642 805 329 413
Field guns 380 1,200 456 800 304
Tanks 78 8,000 624 3,700 289
Combat aircraft 85 9,500 808 5,900 502
2,812 1,596

Table 4 really brings the relative cost of weapons and the critical role of productivity into sharp relief. To take the latter point first, had productivity stalled at the 1941 level, producing the same number of armaments would have required 2.8 billion hours of work instead of 1.6 billion. This was beyond the capacity of the skilled workforce.

More importantly, modern weapons (tanks and aircraft) were much more expensive than those that had dominated the battlefield in World War I (rifles, machine guns and artillery). Assuming the productivity levels of 1943, producing 10 millions rifles and 1.2 million machine guns took 500 million hours of work, about the same as producing just 85,000 combat aircraft. Tanks were expensive too. Each T-34 required 3,700 hours of work; an order of magnitude greater than a machine gun. What this meant for force structure can be determined by the following calculation.

A Soviet rifle division in July 1943 had around 10,000 men armed with rifles, 200 field guns and mortars, and 50 machine guns or anti-aircraft guns. Producing the weapons for the standard rifle division thus required 266,450 hours of work. A Soviet tank corps in July 1943 had 230 tanks in addition to 10,000 men and 160 guns. Arming this standard armored formation required 1,069,000 hours of work. Given the tightness of the skilled labor constraint, this meant that the Soviets faced a tradeoff of four rifle divisions for each tank corps. The air-ground tradeoff was even more daunting. For the price of arming an entire rifle division, the Soviets could obtain a squadron of just 45 dive bombers. 

What compelled both Hitler and Stalin to maximize the production of combat aircraft and tanks and push for wholesale mechanization was the technological-economic conjuncture. In other words, what forced their hand was the internal logic of mobile air-land warfare. The arrival of large numbers of internal combustion engines to the battlefield introduced unprecedented mobility, tempo and depth to the modern battlefield and thereby summoned the ancient gods of the war of movement. It will be argued in a future post that this transformation was not appreciated in the West and lead to catastrophe in 1940. Moreover, it will be argued, that what was essentially a technological transformation came to be associated with the myth of the German Blitzkrieg, whereas in fact the operational art of mobile air-land warfare attained its highest form not in the Wehrmacht but in the Red Army. But for now we must return to economics.

The Soviet armament miracle of the early-1940s was, of course, dependent on an earlier miracle in heavy industry. Basically, it was impossible to produce tens of thousands of tanks and combat aircraft if one did not have a massive base of industries upstream from war production; ie, heavy industry. This heavy-industrial base did not exist in 1928. It came into being during the First Five Year Plan, when the Soviet Union went on an unprecedented buying spree for Western technology, followed by an even more unprecedented construction spree. Suddenly, thousands of gigantic factories mushroomed over the Soviet Union; many of them in factory towns ‘out east’, beyond the Urals—to give to Soviet Union strategic depth against a possible future attack from the west. But even as late as 1933 it was not in fact clear that the leveraged bet would pay off. For it was not enough to erect gigantic factories. Millions had to learn how to master industrial work. What was required was a productivity miracle; for Stakhanovism to truly succeed.

The smashing success of forced-pace industrialization became evident when the results of the Second Plan came in. Figure 3 displays the economic miracle of the mid-1930s. Soviet per capita GNP grew by as astounding 9.2 percent a year during the Second Plan (1933-1937).


Figure 3. Soviet GNP.

Almost all of this growth was concentrated in industry. Industrial growth averaged 13.4 percent in the same period; 9.2 percent in 1929-1940. But Soviet industrialization was weighted towards heavy industry. The critical machinery and machine tools sector (‘mother machines’) grew at 21.5 percent in 1929-40 and a breathtaking 31.7 percent during the Second Plan.

There is disagreement between scholars on industrial productivity growth during the Second Plan, 1933-1937. Industrial employment grew by 24 percent, from 9.3 million to 11.6 million. Estimates for industrial output growth range from Nutter’s 86 percent to Hodgman’s 116 percent. Putting these two together yields an estimated range of 62-92 percent for output per worker. That translates to an annual rate of productivity growth of 10-14 percent during the Second Five Year Plan. The truth must be closer to the upper bound since per capita output growth (industrial and nonindustrial) averaged 9.2 percent in this period. In either case, there was indeed a productivity miracle. Output per worker grew at double digits through the Second Plan.

Both miracles were necessary prerequisites of Soviet victory against Nazi Germany. They spoke to the success of an alternate modernity; of ‘Stalinism as a civilization’. Yet both were largely missed by Western observers. Even today, beyond a vanishingly small group of historians, Westerners remain unaware of the Soviet achievement. We pretend that dashing Americans saved the world from Nazism; that the Western Allies played a major role in defeating the Wehrmacht; and that Stalin and Hitler are somehow morally equivalent. All of these fictions persist because of the rigidity of the discourse of Western self-congratulation.


Glantz, David M., and Jonathan M. House. When Titans Clashed: How the Red Army Stopped Hitler. University Press of Kansas, 2015.

Davies, Robert William, Mark Harrison, and Stephen G. Wheatcroft, eds. The Economic Transformation of the Soviet Union, 1913-1945. Cambridge University Press, 1994.

Harrison, Mark. Accounting for War: Soviet Production, Employment, and the Defence Burden, 1940-1945. Cambridge University Press, 1996.

Harrison, Mark, ed. The Economics of World War II: Six Great Powers in International Comparison. Cambridge University Press, 1998.



The Special Role of the Dive Bomber in the Soviet-German War


The German dive bomber, Stuka.

When Wehrmacht officers could see the Kremlin’s spires through their binoculars and the rest of the Soviet administration was being evacuated from Moscow, Stalin defiantly decided to stay put. News that Stalin stood firm in Moscow calmed the nerves of the city’s jittery residents. In a radio address to the Soviet people shortly thereafter, Stalin declared: “The war will be won by the side that has an overwhelming preponderance in engine production.” He was right on the money.

What became technologically and economically feasible in the late-1930s, and dominated the battlefield in the Soviet-German war, was a novel kind of warfare. What happened to land warfare in the late-1930s was something very specific: Artillery and armor became dramatically more mobile and broke-free from the two-dimensional surface of the earth.

This wasn’t Ivan Bloch’s battlefield that obtained on the western front in World War I. Even when the deadlock broke, the armies of World War I moved at the speed of marching infantry. By the late-1930s, the large-scale deployment of internal combustion engines brought a radical new mobility to both the tactical and the operational levels of land warfare.

Motorized formations could easily outmaneuver any massed formations that moved at the speed of the foot-borne. Large motorized formations composed of infantry, tanks, and artillery fought a complex dance for primacy on the map. The importance of the operational level of land warfare was considerably enhanced by the new operational mobility of the divisions. They could be moved around a map to exploit the terrain and outmaneuver the enemy’s forces. This opened up a whole new game played by the generals on a map of the operational theater.

The new mobility allowed for operational offensives that have been mistaken for the blitzkrieg. The lighting war is a war strategy—a formula for defeating the enemy—that says: Smash a hole through the enemy’s defenses and launch an all-out attack in the rear to achieve a rapid decisive victory.

The new mobility also admitted sub-strategic (ie, not threatening the very existence of the adversary) operational offensives that looked and sounded like this strategy. So successful operational offensives could destroy entire ‘armies’ (large multi-division operational commands) in a single offensive. But such operational blitzkrieg if you will, could not always transform a war of attrition into one admitting a quick decisive victory. With the strategic blitzkrieg, the whole point is to reach a quick decision.

Put another way, the operational blitzkrieg is a motorized solution to the deadlock of the Western front; the strategic blitzkrieg is a motorized solution to the problem posed by the war of attrition. To the extent that the enemy could be conquered in a single offensive (as turned out to be the case in France) the two are hard to distinguish. But distinguish them we must. For after the German strategic blitzkrieg of 1941 failed, operational blitzkrieg continued to dominate warfighting strategy on both sides through 1945 and no solution to the war of attrition could be found.

[Mearsheimer’s doctoral thesis argued that conventional deterrence is relatively easy when such war solutions seem to be available and difficult otherwise. He’s right. Wars of attrition are certainly unappealing if you think that you have less war potential than the enemy. Even if one has a comfortable margin on the enemy to be confident of ultimate victory, the prospect of considerable attrition exercises its own deterrent.]

An equally if not more radical transformation of the battlefield was that it became three-dimensional. Specifically, it became possible to project accurate firepower from the air onto the ground in support of one’s ground forces.

Bombing sorties on enemy formations are not tactical support. World War II bombers were so inaccurate that multiple sorties by flocks of bombers had to flown over a bridge in order to be relatively confident of having destroyed it. Close air support was thus simply not possible with bombers. And in the absence of precision-strike weapons or the capacity to deliver them, there was indeed only one way to deliver accurate fire to the ground from the air. That solution was for aircraft to dive on their targets and attack them with direct-fire. In order to do this effectively, specialized ground-attack aircraft, namely dive bombers, had to be built.

The Soviets, Germans, and the Japanese built dedicated dive bombers and mastered the art of using them in tandem with ground forces. The Western Allies never quite figured it out properly. Perhaps the latter fact explains why this important piece of the puzzle has so far not received the attention it deserves from students of the Soviet-German war.

The dive bombers had to be mass produced in large numbers to count in the war, so all three powers that mastered dive bombing concentrated on a single design they found to be effective. The German Stuka, the Japanese Sonia, and the Soviet Sturmovik were comparably effective machines.


The Soviet dive bomber, Sturmovik.

The Sturmovik was superior to the Stuka. Even if the Red Army deployed their dive bombers with less effect than the Germans, the Soviet Union’s dive bomber force was the strongest in the world by a wide margin. For the Soviet Union made 40,000 of these machines during the war while the Germans made only 6,500. (The Western Allies together produced a much smaller fleet of two thousand such machines; as did the Japanese.)

Combined arms doctrines seek to exploit the complementary nature of the various arms—infantry, artillery, armor, and aircraft—whereby the strength of one arm is used to compensate for weakness of the other. This required the motorization of infantry and artillery, which was achieved partially but substantially by both the Wehrmacht and the Red Army. Once all these pieces fell into place, there emerged a rather large premium on the relative skill in the deployment of the combined arms at both the tactical and operational levels. This was especially true of mobile air-land warfare.

Gaining air superiority—the job of the fighters—wasn’t enough. Airmen had to master dive bombing; ground commanders had to figure out how to effectively use air support; and most of all, officers had to learn how to ensure effective ground-air coordination. The Germans were the first to master dive bombing but the Soviets figured out the secrets to effective ground-attack early in the war as well. The Sturmovik factory in Moscow had to be relocated east of the Urals in the fall of 1941 when the enemy was at the capital’s door. Production promptly resumed before New Year’s Eve.

Whether they became better than the Germans over the course of the war is an open question. My reading is that the Soviets gained increasing mastery of the art and eventually closed the gap with the Germans. The Red Army’s catch up with the Wehrmacht in this technique was, I think, in line with its general catch up in the tactical and operational art of the combined motorized arms.

Ultimately, what mattered was not just relative skill—which only determined the exchange ratios, and more generally, the rates of attrition—but also the capacity of the two party-states to stomach the attrition and generate more fighting power than the enemy. So the game was rigged from the get-go. For the Soviet Union had much greater capacity for the mass production and effective deployment of the war machines than Germany. They could not only make tanks, self-propelled artillery and dive bombers faster than the Germans could shoot them down; they could make dramatically more of them.

This greater capacity of the Soviet Union did not exist in 1931. It was an achievement of Soviet forced-paced heavy industrialization—which was with considerable strategic foresight heavily-weighted towards motorized armament—that made the communist great power militarily much stronger than its overall economic, or even more specifically industrial strength, would otherwise suggest. But it was more than just the achievement of the communist leadership.

The rapidity of Soviet heavy industrialization was the success of ‘Stalinism as a civilization’. For had the Soviet people not bought into Stalinism’s grand project of ‘building socialism’, as Kotkin brilliantly captured, the elite venture would’ve failed. For how was it possible amid the dislocations and informational deficiencies of the command economy and the distorted incentives of the Soviet system—and indeed the definite costs of terror and famine—to transform a Third World country into a great power of the first rank in less than a decade?

What was required was nothing short of the building of a highly dynamic form of machine civilization—where men and women enthusiastically pitted themselves against the machine and each other to achieve ever higher feats of productivity. How Stalinism made all that possible is a fascinating story that has yet to be written. With Kotkin’s Magnetic Mountain we can catch a glimpse of this civilization in its birth pangs.

The nature of the new kind of warfare was not immediately understood by Western observers. Consequently, the premium it put on a very specific kind of heavy industrial capacity was not appreciated. Similarly obscure was the smashing success of Stalinism.

It was understood that the Soviet Union was rapidly industrializing. But no one had ever pulled off a military-industrial transformation of such astonishing rapidity before and there was no reason to believe that the Soviets could either; especially given that instability—famines, dislocations, inefficiencies and terror—seemed to be endogenous to the Soviet system. In addition, there were definite questions about the morale of the Red Army and the Soviet people. Whence, Western intelligence expected the Soviet Union to collapse under the German onslaught in 1941. Even in 1942, the Western Allies continued to worry about a separate Soviet-German peace. It was only by the Spring of 1943 that US intelligence finally became confident that the communist great power was going to prevail against Germany.

The primacy of conventional three-dimensional warfare of the combined mobile arms did not last for long. With the arrival of large numbers of tactical nuclear weapons in the 1950s, massed motorized formations lost their battlefield supremacy. But given the nuclear taboo and limited proliferation, this form of conventional warfare continued to dominate 20th century battlefields and the great powers’ plans to fight conventional war against each other. During the 1980s the rise of reconnaissance-strike ended the relevance of the art of combined mobile arms for conventional warfighting between great powers.