Engels notes that from this point to the end of the chapter the text is taken from Marx's Manuscript VIII.

Volume 1 showed how accumulation works for a single capitalist. When he turns his commodity-capital into money, the surplus product — the part that carries the surplus-value — gets turned into money right along with it. The capitalist then turns this money form of the surplus-value back into extra material elements of his productive capital. In the next round of production, his now-enlarged capital yields an enlarged product.

But what holds for a single capital must also show up in the annual total reproduction of society as a whole — just as we saw with simple reproduction. There, a single capital gradually sets aside, in money, the value of its fixed capital as that capital wears out, and hoards that money; and this too has to show up in society's annual reproduction taken as a whole.

Say a single capital is 400c + 100v, and its annual surplus-value is 100. Then its commodity product is 400c + 100v + 100s. This 600 is turned into money. Of that money, 400 goes back into buying the material form of constant capital, 100 goes to buying labour-power, and — if the whole surplus-value is accumulated — the other 100 is converted into extra constant capital, by buying more material elements of productive capital.

This assumes, first, that under the given technical conditions this sum is actually enough — either to expand the constant capital already at work, or to set up a new business. But it could also be that turning surplus-value into money, and hoarding that money, has to go on for much longer before this can happen at all — before real accumulation, an actual expansion of production, can take place.

Second, it assumes that production on an enlarged scale has already actually begun somewhere else. Because to turn the money — the hoarded surplus-value — into elements of productive capital, those elements have to already be buyable as commodities on the market; and it makes no difference if they aren't bought ready-made but made to order. They are only paid for once they exist, and in any case only after real reproduction on an enlarged scale — an expansion beyond the previous normal level of production — has already taken place for them. They had to be there potentially, that is, in their elements, since all it takes is the trigger of an order — a purchase that comes before the commodity itself exists, an anticipated sale — for their production to actually happen.

So the money on one side calls the enlarged reproduction on the other side into life, because the possibility of that reproduction is already there without the money. Money by itself is not an element of real reproduction at all.

Say capitalist A sells off, bit by bit, the amounts of commodity-product he produces over a year or several years. In doing this he also turns the part of the product that carries the surplus-value — the surplus product — into money bit by bit, and stores it up. This builds him potential new money-capital: potential, because of what it is capable of and meant for — being converted into elements of productive capital.

But what he is actually doing is just simple hoarding, and hoarding on its own is not part of real reproduction. All he is really doing, to begin with, is gradually pulling circulating money out of circulation and locking it away — and this money, before it ever entered circulation, may itself already have been part of somebody else's hoard.

This hoard of A's, potential new money-capital though it is, is not additional wealth for society — no more than it would be if he had spent it on means of consumption instead. Money withdrawn from circulation was, after all, already in circulation before: it might have already sat as part of some other hoard, or been wages in money form, or the proceeds from selling means of production or some other commodity, or have circulated as somebody's constant capital or as a capitalist's revenue.

It is no more new wealth than money — looked at from the standpoint of simple commodity circulation — is a bearer of ten times its own value just because it changed hands ten times in a day and realized ten different commodity-values. The commodities are there without it, and the money itself stays exactly what it is — or gets a little smaller through wear — whether it changes hands once or ten times.

Only in gold production — to the extent that the gold produced includes a surplus product, a carrier of surplus-value — is new wealth (potential money) actually created. And only to the extent that the whole of this new gold product enters circulation does it add to the stock of money-material available for potential new money-capitals.

This surplus-value, hoarded in money form, is not extra new wealth for society — but it does count as new potential money-capital, because of the function it is being stored up for. (We will see later that new money-capital can also arise in other ways besides the gradual turning of surplus-value into gold and silver.)

Money gets pulled out of circulation and piled up as a hoard by selling a commodity without buying anything afterward. Now picture this happening everywhere at once — and it has to be pictured that way, since any single capital at all can be in the middle of accumulating. Then it seems impossible to see where the buyers are supposed to come from: everyone wants to sell in order to hoard, and nobody wants to buy.

Suppose you pictured the circulation between the different parts of the annual reproduction as running in a straight line. That picture is wrong: with only a few exceptions, this circulation always consists of movements running back against each other. But on that false picture, you would have to start with the gold- (or silver-) producer, who buys without ever selling, and assume that everyone else sells to him.

Then the whole of society's annual surplus product — the carrier of the whole surplus-value — would pass over to him, and every other capitalist would share his surplus product out among themselves, each taking a share in proportion to its own surplus-value, since that product already exists by its very nature in the form of money: the natural gold-form of his surplus-value. (The part of the gold producer's own product that has to replace the capital he already has at work is already spoken for.) The surplus-value the gold producer produces in gold would then be the one and only fund that every other capitalist draws on to turn their own annual surplus product into money. It would have to equal, in value, the whole of society's annual surplus-value — which would first have to cocoon itself into the form of a hoard.

However absurd these assumptions are, all they could do is explain how a general, simultaneous piling-up of hoards is possible at all. Reproduction itself would be no further advanced by any of it — except on the gold producers' side.

Before we resolve this apparent difficulty, we need to distinguish between accumulation in department I, which produces means of production, and accumulation in department II, which produces means of consumption. We will start with department I.

Think about all the different businesses that make up Department I — many industries, and within each industry many individual firms. They differ in age: how long each one has already been running. Set aside their size, their technical setup, how their markets are doing — none of that matters here. What matters is that each firm is at some different point in a process: turning its surplus-value, step by step, into money-capital that hasn't been put to work yet. That money-capital, once it exists, can go two ways — it can be added to enlarge the capital a firm already has running, or it can go toward setting up an entirely new business. Those are the two ways production can expand.

So at any moment, some capitalists have already built up enough of this stored-up money and are now converting it into productive capital: they take the money they saved from selling their surplus product and use it to buy means of production — extra buildings, machines, materials, whatever adds to their constant capital. Other capitalists are still at the earlier stage, still building up their stock of money and not yet spending it.

This puts the two groups face to face: one group as buyers, the other as sellers — each one stuck in that single, exclusive role.

Say A sells 600 worth of goods (=400c+100v+100s) to B — who might stand in for more than one buyer. A has sold 600 in commodities for 600 in money, and 100 of that money represents surplus-value, which he pulls out of circulation and hoards as money. But that 100 in money is nothing more than the money-form of the surplus product — the actual goods — that were worth 100 to begin with.

Hoarding like this is not production at all — so it can never be a further increment of production either. All the capitalist is doing here is pulling the money he got from selling that 100 worth of surplus product out of circulation, holding onto it, sitting on it. And this isn't just A: the same thing is happening at countless other points around the circuit, with other capitalists — call them A′, A″, A‴ — all just as busy building up hoards of their own.

All these countless points, where money gets pulled out of circulation and piles up into separate hoards, into potential money-capital sitting idle, look like so many roadblocks in the way of circulation — they freeze the money and stop it circulating for a longer or shorter stretch. But consider: this kind of hoarding already happens in plain commodity circulation, long before that circulation is built on capitalist commodity production. The amount of money present in a society is always bigger than the part of it actually circulating at any moment, even though that active part grows and shrinks with circumstances. We are looking at these very same hoards and this very same hoarding here — except now they are a moment built into the capitalist production process itself.

It's easy to see the appeal: once the credit system is in place, banks and the like gather up all these separate stores of potential capital and turn them into capital that's available to lend out — 'loanable capital', money-capital. That turns it from something passive, a mere promise for later, into something active and multiplying — 'multiplying' here just in the sense of growing, not of squeezing out interest.

A only manages to build up this hoard on one condition: that when it comes to his surplus product, he acts purely as a seller, never afterward as a buyer. His hoarding depends on this — it requires that he keep producing surplus product, round after round, since that surplus product is what carries the surplus-value he still has to turn into money.

In the case we're looking at, where we're only tracking circulation within Department I, the actual, physical form of this surplus product — like the physical form of the whole product it's part of — is some element of constant capital for Department I. That is, it belongs to the category of means of production used to make other means of production. What becomes of it, what job it ends up doing once it's in the hands of the buyers B, B′, B″ and so on — we'll see that shortly.

Here's the key thing to hold onto: A pulls money out of circulation for his surplus-value and hoards it — but at the same time, he throws commodities into circulation without taking any other commodities back out in exchange. That one-sided move is exactly what lets B, B′, B″ and the others throw money into circulation and take out only commodities, without putting any commodities in themselves.

In the case here, this commodity — both by its physical form and by what it's for — becomes part of B's (B′'s, and so on) constant capital, either as a fixed element or as a circulating one. More on that once we turn to the buyer of the surplus product, B, B′, and the rest.

One thing worth noting in passing: just as before, when we were looking at simple reproduction, we find here too that exchanging the different parts of the year's product — that is, their circulation, which has to include the reproduction of capital in all its various forms (constant, variable, fixed, circulating, money-capital, commodity-capital) — does not simply mean a purchase of goods that gets completed later by a matching sale, or a sale completed later by a matching purchase, as if the whole thing came down to trading goods for goods. That is what political economy assumes — especially the free-trade school, going back through Adam Smith to the Physiocrats.

We already know that fixed capital, once the outlay for it has been made, is not renewed for its entire working life; it goes on working in its old physical form the whole time, while its value gradually settles into money bit by bit. We saw that the periodic renewal of fixed capital in IIc — where the whole capital-value of IIc converts into elements worth I(v+s) — requires two things at once: on one side, a plain purchase by the fixed part of IIc, changing back from money-form into its physical form, matched by a plain sale from Is, department I's surplus-value share; on the other side, a plain sale by IIc — selling off the worn portion of its fixed capital's value, which settles into money — matched by a plain purchase from Is.

For this exchange to go normally, it has to be the case that IIc's plain purchases equal, in value, IIc's plain sales; and likewise, that the plain sale from Is to the first part of IIc equals, in value, its plain purchase from the second part of IIc. If not, simple reproduction is thrown off course — a plain purchase on one side must be matched by a plain sale on the other. In just the same way, it has to be the case here that the plain sale made by A, A′, A″ — the ones building up hoards — out of their share of Is, balances against the plain purchase made by B, B′, B″ — the ones turning their hoard into elements of additional productive capital.

Insofar as balance comes about because the buyer later turns around and sells the same amount of value, and the seller later turns around and buys the same amount, money flows back to whichever side advanced it in the first purchase — the side that sold before it bought again. But the real balance, the one that actually matters for the exchange of goods itself, for exchanging the different parts of the year's product, depends on the goods traded against each other being equal in value.

But insofar as the exchanges are purely one-sided — a mass of plain purchases on one side, a mass of plain sales on the other — and we've already seen that normal exchange of the year's product, on a capitalist basis, requires exactly this kind of one-sided movement — balance only exists on one condition: that the total value of the one-sided purchases matches the total value of the one-sided sales.

The fact that commodity production is the general form capitalist production takes already brings with it the role money plays in it — not just as a means of circulation, but as money-capital. And that generates certain conditions, particular to this way of producing, for normal exchange to happen — that is, for reproduction to run its normal course, whether on the same scale or on an enlarged one.

But those very same conditions turn, in the same movement, into conditions for an abnormal course — into possibilities of crisis. Because under this kind of production, which grows up on its own rather than being planned, balance itself is a matter of chance.

We've also seen that in the exchange of Iv — the wages part of department I's product — against the matching value in IIc, the constant-capital part of department II's, what happens for IIc in the end is this: commodity II gets replaced by an equal value of commodity I — in other words, the capitalists of Department II, having sold their own goods, later complete the operation by buying commodity I for the same amount. This replacement really does happen. But it isn't a direct exchange between the capitalists of I and II swapping their goods with each other.

Here's how it actually runs: IIc sells its goods to the working class of Department I. The workers face IIc purely as buyers of goods; IIc faces them purely as a seller of goods. With the money IIc gets this way, IIc then faces the capitalists of Department I purely as a buyer of goods — and up to the amount of Iv, the capitalists of I face IIc purely as sellers of goods. It's only through this sale that Department I finally gets its variable capital back in the form of money.

So: the capital of Department I faces Department II purely as a seller of goods, up to the amount of Iv — and faces its own working class purely as a buyer, buying their labour-power. And the working class of Department I faces capitalist II purely as a buyer of goods, buying means of subsistence — and faces capitalist I purely as a seller of goods, namely as the seller of its own labour-power.

The working class of Department I has to keep on offering its labour-power, without a break. Part of commodity-capital I has to turn back into money-form as variable capital. Part of commodity-capital II has to be replaced by the physical elements that make up constant capital IIc. All three of these are necessary conditions, each depending on the other two — but they don't happen directly. They're carried out through a very complicated process, made up of three circulation processes - those three - that run independently of each other yet are tangled together. And the sheer complicatedness of this process is itself just as many chances for things to go wrong.

The surplus product — the thing that carries the surplus value — costs capitalists I nothing to get. They don't have to lay out any money or commodities in advance to obtain it. ("Advance" has meant, since the Physiocrats, value laid out and turned into elements of productive capital.) All they advance is their constant and variable capital. The worker gives them back their constant capital through his labour; he also replaces the value of their variable capital with a newly created equal value, in the form of a commodity. And beyond that, through his surplus labour, he hands them a surplus value existing in the form of a surplus product. By selling this surplus product bit by bit, the capitalists build up a hoard: additional money capital, so far only potential.

In the case we're looking at, this surplus product consists, from the outset, of means of production for making other means of production. It only starts working as additional constant capital once it reaches the hands of B, B′, B″ and the rest (in department I). But it already is this, in potential, before it's even sold — already in the hands of A, A′, A″, the ones building up the hoard.

If we look only at the sheer amount of value being reproduced by department I, we're still inside the bounds of simple reproduction: no extra capital was set in motion to create this potentially-additional constant capital (the surplus product), and no more surplus labour was spent than simple reproduction already required. The only difference is in the form the surplus labour took — the particular useful shape it was poured into. It went into means of production for Ic rather than for IIc; into means of production for making means of production, rather than means of production for making means of consumption. Under simple reproduction, the assumption was that the whole of surplus value I gets spent as revenue — that is, on commodities from department II — so it consisted only of means of production suited to replacing constant capital IIc in its own natural form.

So for the shift from simple to expanded reproduction to happen at all, production in department I has to be able to turn out fewer elements of constant capital for II, and correspondingly more for I. This shift doesn't always come easily, but it's made easier by the fact that a good number of department I's products can serve as means of production in either department.

So it follows that — looking only at value-scope — the physical basis for expanded reproduction gets produced within simple reproduction itself. It's nothing more than the surplus labour of department I's working class, spent directly on producing means of production, and so creating potential additional capital for department I. When A, A′, A″ and the rest build up potential additional money capital — by selling off their surplus product bit by bit, a surplus product they got without laying out any capitalist money at all — that money capital is just the money-form of the extra means of production department I has already produced.

Producing potential additional capital, then — in the case here (though as we'll see, it can also arise in a completely different way) — is nothing but a phenomenon of the production process itself: production, in one particular form, of elements of productive capital.

So when we see potential additional money capital being produced on a large scale, at many points around the edge of circulation, this is nothing but the result and expression of many-sided production of potentially additional productive capital — capital whose creation required no additional money outlay from the industrial capitalists at all.

This potentially additional productive capital gets turned, bit by bit, into potential money capital — a hoard — by A, A′, A″ and the rest. That happens because they keep selling their surplus product one-sidedly, without buying anything in return, and each such sale pulls money out of circulation and adds it to the growing hoard.

Except in one case — where the buyer is the gold producer, paying with newly mined gold — this hoard-building doesn't require any additional wealth in precious metal at all. It only requires a change in the function of money that was already circulating. A moment ago that money was serving as a means of circulation; now it serves as a hoard, as newly forming potential money capital. So the formation of additional money capital and the total mass of precious metal sitting in a country have no causal connection to each other.

It follows, further: the bigger the productive capital already at work in a country (counting the labour-power built into it — the labour-power that produces the surplus product) — the more developed the productive power of labour, and with it the technical means for rapidly expanding the production of means of production — and so the bigger the mass of the surplus product, both in value and in the mass of use-values it takes the form of — the bigger, then, is:

1. the potential additional productive capital sitting as surplus product in the hands of A, A′, A″ and the rest, and

2. the mass of the surplus product once it's turned into money — that is, the potential additional money capital in the hands of A, A′, A″.

So when someone like Fullarton says he wants nothing to do with overproduction in the ordinary sense, but is happy to talk about overproduction of capital — meaning overproduction of money capital — that just proves how little even the best bourgeois economists understand the mechanism of their own system.

Here's the thing: the surplus product — produced and taken over directly by capitalists A, A′, A″ (I) — is the real basis of capital accumulation, that is, of expanded reproduction. And yet it only actually functions in that role once it's in the hands of B, B′, B″ and the rest (I).

In its money disguise, though — as a hoard, as potential money capital only gradually taking shape — it's a different story: this form is completely unproductive. It runs alongside the production process without being part of it; it lies outside it. It is dead weight on capitalist production.

The urge to put this surplus value — piling up as potential money capital — to work, both for profit and as revenue, is what drives people toward the credit system and its little paper securities. Through that route, money capital gains, in a different form, enormous influence over the course and the vast development of the capitalist system of production.

The bigger the total capital already at work — the capital whose functioning gave rise to this potential money capital — the bigger the mass of surplus product converted into potential money capital will be.

But as the yearly-reproduced potential money capital grows in absolute size, it also becomes easier to split up. That means it gets invested faster in some particular business of its own — whether by the same capitalist, or by other people (family members dividing an inheritance, say). "Splitting up" money capital, here, means separating it completely from the parent capital, so it can be invested as new money capital in a new, independent business.

The sellers of the surplus product — A, A′, A″ and the rest (I) — got it as the direct result of the production process itself, a process that, beyond the same advance of constant and variable capital simple reproduction already required, needs no further act of circulation at all. In supplying it, they're delivering the real basis for reproduction on an expanded scale — in fact, they're manufacturing potential additional capital.

B, B′, B″ and the rest (I) are in a different position, though, in two ways. First, it's only once the surplus product reaches their hands that it actually starts functioning as additional constant capital. (We're leaving aside, for now, the other piece of productive capital — the additional labour-power, that is, the additional variable capital.) Second, for it to reach their hands at all, an act of circulation is needed: they have to buy it.

On the first point: a large part of this surplus product — potential additional constant capital, produced by A, A′, A″ (I) — does get produced this year, but can only actually start functioning as industrial capital in the hands of B, B′, B″ (I) next year, or even later.

On the second point, the question is: where does the money needed for this act of circulation come from?

Now, to the extent that what B, B′, B″ and the rest (I) themselves produce feeds straight back into their own process, in kind, it's obvious that part of their own surplus product passes directly — with no need for circulation — into their productive capital, entering it as an extra element of constant capital. But to that same extent, they aren't the ones turning A, A′ and the rest's (I) surplus product into money either.

Setting that aside, then — where does the money come from? We already know they built up their own hoard the same way A, A′ and the rest did: by selling their respective surplus products. And now they've reached the point where that hoarded, still-only-potential money capital is supposed to start actually functioning as additional money capital.

But that just takes us in a circle. The question is still exactly where the money came from that the B's (I) withdrew from circulation and piled up in the first place.

But we already know, from looking at simple reproduction, that a certain amount of money has to sit in the hands of capitalists I and II to turn their surplus product into cash. There, the money — spent only as revenue on means of consumption — flowed back to the capitalists in step with how much they'd advanced to sell their own commodities. Here, that same money turns up again, but doing a different job.

The A's and the B's (I) take turns supplying each other with the money needed to convert surplus product into additional potential money capital — and take turns throwing the newly formed money capital back into circulation as a means of purchase.

The only thing this assumes is that the amount of money in the country — taking the speed it circulates at, and so on, as fixed — is enough to cover both active circulation and the reserve hoard together. That's the very same condition we saw has to hold even for simple commodity circulation; only the job the hoards do is different here.

The money on hand also has to be bigger than before, for four reasons. First, under capitalist production almost everything gets produced as a commodity — except newly-mined precious metal and the small amount producers consume themselves — so almost everything has to pass through a money-disguise at some point. Second, on a capitalist footing the mass of commodity capital, and its total value, isn't just bigger outright — it grows far faster than before. Third, an ever-larger variable capital constantly has to be converted into money capital. Fourth, as production expands, new money capitals keep forming to match it, so the raw material for their hoard-form has to be on hand too.

This holds without qualification in the first phase of capitalist production, where the credit system still runs mostly alongside metallic circulation. But it holds even in the most developed phase of the credit system, so far as that system's basis remains metallic circulation. On one side, extra production of precious metals — when it swings between plentiful and scarce — can disturb commodity prices, and not just over long stretches but within very short ones too. On the other side, the whole credit mechanism is constantly busy squeezing actual metal circulation down toward an ever-shrinking minimum, through every kind of operation, method, and technical device — and as it does, the whole apparatus gets more artificial, and the chances of something disrupting its normal course grow right along with it.

The various B's — B, B′, B″ and the rest (I) — whose potential new money capital has now become active, may well need to buy from and sell to each other: parts of their own surplus product changing hands among them. To that extent, the money advanced to circulate the surplus product flows back — in the normal case — to the various B's, in the same proportion each of them advanced it to circulate their own commodities. If the money is circulating as a means of payment, then only the balances need to be settled, wherever these mutual purchases and sales don't exactly cancel out.

But it matters — here as everywhere in this account — to start by assuming metallic circulation in its simplest, most original form. That way, flow and reflux, the settling of balances, and in short everything that shows up in the credit system as a consciously managed process, can be seen existing independently of the credit system — the whole thing appearing in its naturally grown shape, rather than in the later, more self-aware one.

So far this has all been about additional constant capital. Now we need to turn to additional variable capital.

Volume 1 explained at length how, under capitalist production, labour-power is always available in reserve, and how — when it's needed — more labour can be squeezed out without hiring more workers or drawing on more labour-power. There's no need to go over that again here for the moment; we can just assume that whatever part of the newly-formed money capital is convertible into variable capital will always find the labour-power to convert it into.

Volume 1 also explained how a given capital, without any accumulation at all, can expand how much it produces, within certain limits. But here we're dealing with capital accumulation in the specific sense: production expands only because surplus value gets converted into additional capital — which means an expanded capital-basis for production too.

The gold producer can accumulate part of his own gold surplus value directly as potential money capital. Once it reaches the size he needs, he can turn it straight into new variable capital, without first having to sell any surplus product at all. He can do the same to turn it into elements of constant capital.

But in that second case, he still has to find the actual physical elements of his constant capital available. That might mean, as we've been assuming so far, that every producer works to build up stock and then brings the finished commodity to market — or it might mean he works to order. Either way, a real expansion of production — that is, a surplus product — is presupposed: in one case actually there already, in the other only potentially there, ready to be delivered.

So far we've assumed that A, A′, A″ in department I sell their surplus product to B, B′, B″ — capitalists who belong to that same department I. But now suppose instead that A in department I turns his surplus product into money by selling it to a B in department II. That can only happen one way: A sells B means of production, and afterward does not buy means of consumption back from him. In other words, only through a sale that runs one way, from A's side alone.

Here is why that matters. IIc can only convert back from commodity capital into the natural form of productive constant capital if not just Iv, but also at least part of Is — department I's surplus product — gets exchanged for a part of IIc — the part that exists as means of consumption. But now A turns his surplus product into money precisely by not completing that exchange. Instead of using the money from his sale to buy means of consumption from B, he pulls it out of circulation and holds onto it.

So on A's side, this does produce additional virtual money capital. But on the other side, an equal amount of value sits frozen in B's constant capital — stuck in the form of unsold commodities, unable to convert back into the natural form of productive constant capital. In other words: part of B's goods — at first glance, exactly the part he needs to sell in order to fully turn his constant capital back into productive form — has become unsellable. Overproduction has occurred with respect to that part. And that same part is holding back reproduction — even reproduction on the same scale as before.

So in this case, A's additional virtual money capital is indeed the money-form of surplus product — of surplus-value. But surplus product, surplus-value, considered just as such, is here still a phenomenon of simple reproduction. It is not yet reproduction on an expanded scale. I(v+s) — or at least, of the surplus part, however much of it the requirement reaches — has to end up being exchanged against IIc, or IIc cannot reproduce on the same scale as before.

By selling his surplus product to B, A has delivered him a matching share of constant capital in its natural form. But at the same time, by pulling the money out of circulation — by never completing his sale with a follow-up purchase — he has made an equal-value share of B's goods unsellable.

So look at social reproduction as a whole, taking in capitalists I and II together. A's surplus product turning into virtual money capital is really just the flip side of an equal amount of B's commodity capital failing to convert back into productive constant capital. This is not, even virtually, production on an expanded scale. It is a hampering of simple reproduction — a deficit in simple reproduction itself.

Since A producing and selling his surplus product are themselves perfectly normal features of simple reproduction, what we have here — on the ground of simple reproduction alone — is a set of phenomena that all depend on one another: virtual additional money capital forming in class I, which means underconsumption seen from class II's side; commodity stocks piling up in class II that cannot convert back into productive capital, which is relative overproduction in II; surplus money capital in I, and a deficit in reproduction in II.

Without dwelling on this point any further, one thing is worth noting. In laying out simple reproduction, we assumed that the whole of surplus-value in I and II gets spent as revenue. In reality, though, part of surplus-value gets spent as revenue and another part gets turned into capital. Real accumulation only happens on that basis.

The idea that accumulation happens at the expense of consumption, stated in such general terms, is itself an illusion — one that contradicts the very nature of capitalist production. It assumes that the purpose and driving motive of capitalist production is consumption, when it is really the grabbing of surplus-value and turning it into capital, that is, accumulation.

Now let's take a closer look at accumulation in department II.

The first difficulty concerning IIc — that is, converting it back from being part of department II's commodity capital into the natural form of department II's constant capital — belongs to simple reproduction itself. Let's take the earlier schema:

(1,000v + 1,000s) I exchange against:

2,000 IIc.

Suppose now that half of I's surplus product — 1,000/2 s, that is, 500 Is — is earmarked to function as additional constant capital within department I itself, instead of going to department II. Then this portion, kept back within I, cannot replace any part of IIc. It was supposed to be converted into means of consumption — and this piece of circulation between I and II is a genuine two-way trade, goods actually changing hands on both sides, unlike the replacement of 1,000 IIc by 1,000 Iv, which runs through the workers' spending. Instead, it is meant to serve as an additional means of production within I itself. It cannot do both at once. The capitalist cannot spend the value of his surplus product on means of consumption and, at the same time, productively consume that very surplus product himself by building it into his own productive capital.

So instead of the full 2,000 I(v+s), only 1,500 — that is, (1,000v + 500s) I — can be exchanged against the 2,000 IIc. Which means 500 of II's goods cannot convert back out of commodity form into productive constant capital II. Overproduction would then have taken place in II, matching exactly the extent of the expansion that took place in I. This overproduction in II might react back on I so strongly that even the 1,000 that I's workers spent on means of consumption from II would only partly flow back — meaning that money would not fully return, as variable money capital, into the hands of capitalists I. Those capitalists would then find themselves held back even in reproduction on the same scale as before — and by nothing more than the mere attempt to expand it.

But weigh this: in I, only simple reproduction actually took place. Nothing was really added. All that happened is that the very same elements shown in the schema got grouped differently, for the sake of an expansion still to come — say, next year.

One might try to sidestep the difficulty by objecting as follows. The 500 IIc sitting in the capitalists' stock, not directly convertible into productive capital, are so far from being overproduction that they are, on the contrary, a necessary element of reproduction — one we have so far left out of account. We saw earlier that money piles up at many points and has to be pulled out of circulation: partly to allow new money capital to form within I itself, partly to hold, for the time being, the value of fixed capital as it slowly wears away, in money form.

But in this schema, all the money and all the commodities are, from the start, exclusively in the hands of capitalists I and II. There is no merchant here, no money-dealer, no banker, no class that merely consumes without taking part directly in commodity production. So the constant build-up of commodity stocks — here, in the hands of the very producers who hold them — is just as indispensable here as that money-stock was, if the machinery of reproduction is to keep running.

The 500 IIc sitting in the stock of capitalists II, then, represent the stock of means of consumption that carries the consumption process built into reproduction from one year over into the next. This consumption fund, still sitting in the hands of its own sellers and producers, cannot sink to zero this year only to start again at zero next year — no more than that could happen going from today into tomorrow. Since such stocks must constantly be renewed, even if their size varies, our capitalist producers in II must have a reserve of money capital that lets them keep production going even while part of their productive capital is temporarily tied up in commodity form. After all, by assumption, they combine the whole business of merchant and producer in one; so they must also have on hand the additional money capital that, once the different functions of the reproduction process split off among different kinds of capitalists, ends up sitting in the hands of merchants.

The reply has three parts.

First: this kind of stock-building, and its necessity, holds for all capitalists, both I and II. As mere sellers of commodities, they differ only in which kind of goods they sell. A stock of commodities in II presupposes an earlier stock of commodities in I. If we leave this stock out of account on one side, we have to leave it out on the other side too. And if we take it into account on both sides, nothing about the problem changes.

Second: just as this year ends, on II's side, with a commodity stock left over for next year, so it also began with a commodity stock on that same side, handed down from last year. In analysing annual reproduction — reduced to its plainest terms — we have to cancel this stock out both times. We let this year keep its whole output, including what it hands over as stock to next year — but we also take away, on the other side, the stock it received from last year. What is left is simply the total product of an average year, which is what we are actually analysing.

Third — and this is the real point — the simple fact that this difficulty we are trying to sidestep never came up while we were looking at simple reproduction proves that it arises from the changed grouping of department I's elements, and from nothing else. It is owed only to a changed grouping of I's elements, for the purposes of reproduction — a changed grouping without which reproduction on an expanded scale could not take place at all.

Let's now look at reproduction using the following schema:

The first thing to notice: the year's total social product comes to 8,252 — smaller than the 9,000 in the earlier schema. A much bigger sum would have worked just as well; it could have been ten times as large, for all the difference that makes. A smaller sum than the earlier schema's was picked on purpose, to make one thing plain: reproduction on an expanded scale — understood here simply as production carried on with a bigger outlay of capital — has nothing to do with the sheer size of the product. For a given mass of commodities, it calls for nothing more than a different arrangement, a different assignment of jobs, among that same product's existing elements. So, measured by value, it is at first nothing but simple reproduction.

What changes is not the quantity of the elements already present in simple reproduction, but their qualitative role — which job each one does. And this change of role is the material precondition for the reproduction on an expanded scale that follows later.

We could set out the schema differently, too, with a different ratio between variable and constant capital — like this, for instance:

Schema (b) would look, on the face of it, set up for reproduction on the same scale as before — its surplus-value spent entirely as revenue, none of it accumulated.

Either way — schema (a) or schema (b) — we have an annual product of the same total value. The only difference is how its pieces are grouped by function. Under (b), that grouping starts reproduction over again at the same scale. Under (a), it forms the material basis for reproduction on an expanded scale instead.

Specifically: under (b), (875v + 875s) of department I — 1,750 I(v+s) — exchanges evenly against 1,750 IIc, with nothing left over. Under (a), (1,000v + 1,000s) of department I — 2,000 I(v+s) — exchanges against only 1,500 IIc, leaving a surplus of 500 Is over for accumulation in department I.

Now for a closer look at schema (a). Suppose that in both department I and department II, half the surplus-value gets accumulated instead of spent as revenue — turned into an element of additional capital.

Since half of the 1,000 in department I's surplus-value — 500 — is to be accumulated one way or another, laid out as additional money-capital and so turned into additional productive capital, only 1,000v + 500s of department I gets spent as revenue. So the normal size of IIc here comes to only 1,500 as well.

The exchange between 1,500 I(v+s) and 1,500 IIc needs no separate examination — it's already been set out as a process of simple reproduction. The same goes for department I's existing constant capital, 4,000 Ic: how it gets rearranged for the new round of reproduction — this time on an expanded scale — was also covered as a process of simple reproduction.

So what's left to examine is only this: the 500 Is left over in department I, and the (376v + 376s) of department II — both their internal makeup and the movement between the two.

Since department II, like department I, is assumed to accumulate half its surplus-value, that means turning 188 into capital here. Of that, a quarter goes to variable capital — 47, or, to round it off, 48. That leaves 140 to be turned into constant capital.

Here we run into a new problem — one whose sheer existence must look strange, given the ordinary understanding that goods of one kind get exchanged for goods of another kind, and likewise goods for money, and that same money again for goods of some other kind.

The 140 of department II's surplus-value can only be turned into productive capital if it's replaced by a portion of department I's goods worth the same amount. It goes without saying that whatever part of department I's goods gets exchanged for it must consist of means of production — the kind that can go into production either in both departments, or in department II alone.

This exchange can only happen through a one-sided purchase by department II. Why one-sided? Because the whole of the remaining surplus product still to be considered — the 500 in department I's surplus-value — is earmarked for accumulation inside department I itself, so it can't be exchanged for department II's goods: department I cannot both accumulate that surplus product and eat it at the same time. So department II has to buy that 140 with hard cash — cash that doesn't flow back through any later sale of department II's goods to department I.

And this is a process that keeps repeating, every year, with every fresh round of production, for as long as reproduction is happening on an expanded scale. So where, in department II, does the money for this come from?

Department II looks, on the contrary, like thoroughly barren ground for forming new money-capital — the kind of money-capital that accompanies real accumulation and, under capitalist production, precedes it, even though in practice it first shows up as nothing more than plain hoarding.

Start with the 376 that is department II's variable capital. This 376 in money-capital, advanced to pay for labour-power, keeps coming back to the department II capitalists as variable capital in money form, through purchases of department II's own goods. This constant movement away from and back to its starting point — the capitalist's pocket — doesn't increase, in any way, the money circulating around this loop. So this is no source of money-accumulation. Nor can this money be pulled out of that circulation to build up a hoard — a potential new money-capital.

But wait a minute — isn't there a little profit to be made here?

We shouldn't forget that department II has an advantage department I doesn't: the workers it employs have to buy back, from it, the very goods those workers produced themselves. Department II is both the buyer of labour-power and, at the same time, the seller of goods to the very people whose labour-power it bought. So here is what department II can do:

(1) One thing department II could do — and this it shares with department I's capitalists — is simply push wages below their normal average. That frees up part of the money that was functioning as the money-form of variable capital — the money laid out on wages, and if the same move were repeated over and over, it could become a normal source of hoard-building — and so, of forming virtually additional money-capital in department II.

We're not talking here about some occasional swindle-profit; this is meant to explain normal capital-formation. But it must not be forgotten: the wage actually, normally paid — which, other things being equal, fixes the size of variable capital — is not handed over out of the capitalists' generosity. It has to be paid, given the conditions capitalists actually face. That rules this explanation out. Having assumed 376v as the variable capital department II lays out, we cannot — just to explain a newly-arisen problem — suddenly smuggle in the assumption that it really only advances 350v, not 376v.

(2) On the other hand, department II as a whole has, as already said, an advantage over department I: it is both the buyer of labour-power and the seller who resells its own goods back to those same workers. And how that can be exploited — how the normal wage can be paid in name only, while part of it gets snatched back without any equivalent in goods to show for it, whether through a company-store arrangement or by tampering with the circulating currency, even where that tampering can't quite be pinned down as illegal — the plainest evidence for this exists in every industrial country, England and the United States among them.

But this is the very same operation as the first one, just dressed up and carried out the long way round. So it has to be rejected here exactly as that one was. What is at issue is wages really paid — not wages nominally promised.

So we can see: an honest, objective analysis of how the capitalist mechanism actually works cannot use certain shameful practices — ones that still cling to it with remarkable persistence — as an excuse for dodging theoretical difficulties.

But oddly enough, most of my bourgeois critics complain that I do the capitalist an injustice — by assuming, in Volume One of Capital for instance, that he pays the real value of labour-power, which in most cases he doesn't! (Schäffle can be quoted here, on the magnanimity he ascribes to me.)

So the 376 that is department II's variable capital gets us no nearer to the goal we've been discussing.

But things look even more doubtful with the 376 that is department II's surplus-value. Here only capitalists of the same department face each other, selling to and buying from one another the means of consumption they themselves produced. The money this exchange needs functions only as a means of circulation, and — in the normal course of things — has to flow back to whoever advanced it, in proportion to what each put in, so it can run the same circuit over again.

Pulling department II's surplus-value money out of circulation this way, to form virtually additional money-capital, seems possible only in two ways.

First: some of department II's capitalists could swindle the others, robbing them of their money. Forming new money-capital, as we already know, doesn't need any prior increase in the money supply — all it needs is money withdrawn from circulation at certain points and piled up as a hoard. That the money involved might be stolen — so that one group of department II's capitalists builds up additional money-capital while another group takes an actual loss — has no bearing on the point being made here. The swindled capitalists would just have to live a bit less extravagantly. That's all there is to it.

Or else: a part of department II's surplus-value — the part that exists as necessary means of subsistence — gets turned directly into new variable capital within department II itself. How this happens will be examined at the end of this chapter, in section 4.

Accumulation

Assume that in this version, department I sets aside half its surplus value to accumulate — that's 500. First we get 1,000 in variable capital plus 500 in surplus value, 1,500 department I (variable capital plus surplus value) in all, to be exchanged for 1,500 of department II's constant capital. That leaves department I with 4,000 in constant capital plus 500 in surplus value still to be accumulated. Exchanging that 1,500 from department I for department II's 1,500 is simple reproduction — the same process already explained there.

Suppose that of the 500 in surplus value, 400 is to become constant capital and 100 variable capital. How that 400 moves within department I once it's turned into capital has already been worked out: it can simply be added onto department I's constant capital. That gives department I: 4,400 in constant capital, 1,000 in variable capital, and 100 in surplus value still to be turned into variable capital.

For the sake of department I's accumulation, department II buys that 100 — existing as means of production — from department I. It becomes additional constant capital for department II. The 100 in money that department II pays for it becomes, in money form, additional variable capital for department I. Department I's capital is now 4,400 in constant capital plus 1,100 in variable capital (the latter in money) — 5,500 in all.

Department II now has 1,600 in constant capital to work up. To do so it has to lay out a further 50 in money to buy new labour-power, so its variable capital grows from 750 to 800. This whole expansion of constant and variable capital together, 150, has to come out of department II's own surplus value. So of the 750 in surplus value, only 600 is left as the capitalists' fund for their own consumption. Department II's yearly product now breaks down like this:

Department II: 1,600 in constant capital, plus 800 in variable capital, plus 600 as the capitalists' consumption fund — 3,000 in all.

The 150 produced as means of consumption — the goods that get exchanged here for department II's 100 constant capital plus 50 variable capital — go, in natural form, entirely to workers' consumption: 100 eaten by department I's workers, 50 by department II's own, as already explained.

In fact, department II — where its whole product has to be put into the shape accumulation requires — has to reproduce 100 more of its surplus value in the form of necessary means of consumption than it otherwise would. If reproduction on an expanded scale actually gets under way, then the 100 in variable money capital from department I flows back, through the hands of its own working class, to department II — which, in turn, hands over 100 worth of goods in stock to department I, and at the same time 50 worth of goods in stock to its own working class.

Now here's how the arrangement looks, once it has been changed to make room for accumulation:

Here is the capital portion of that new arrangement:

Production, though, actually started the year with:

If real accumulation now goes ahead on this basis — that is, if production is actually carried out with this enlarged capital — then at the end of next year we get:

Now let department I go on accumulating in the same proportion: 550 in surplus value spent as revenue, 550 accumulated. First, the 1,100 in department I's variable capital gets replaced by 1,100 of department II's constant capital; on top of that, a further 550 in department I's surplus value still has to be realized against an equal amount of department II's goods — 1,650 department I (variable capital plus surplus value) in all. But the constant capital department II needs replaced comes only to 1,600, so the remaining 50 has to be made up out of its 800 in surplus value. Setting money aside for the moment, here is the result of this exchange:

Department I: 4,400 in constant capital, plus 550 in surplus value still to be capitalized. Alongside that, 1,650 — variable capital plus surplus value — sits in the capitalists' and workers' consumption fund, realized in department II's goods.

Department II: 1,650 in constant capital (that is, with the 50 just added from its surplus value), plus 800 in variable capital, plus 750 in surplus value as the capitalists' consumption fund.

But if the old ratio of variable capital to constant capital in department II still holds, then a further 25 in variable capital has to be laid out for that 50 in constant capital — and it has to come out of the 750 in surplus value. So we get:

Department II: 1,650 in constant capital, plus 825 in variable capital, plus 725 in surplus value.

In department I, 550 in surplus value is to be capitalized; if the earlier ratio holds, 440 of that forms constant capital and 110 forms variable capital. That 110, in this case, has to be drawn from department II's 725 in surplus value — meaning that means of consumption worth 110 are eaten by department I's workers instead of by department II's capitalists. Those capitalists are then forced to capitalize the 110 they can no longer consume themselves. That leaves 615 out of the 725.

But once department II turns that 110 into additional constant capital this way, it needs a further 55 in additional variable capital — and that, too, has to come out of its surplus value. Deducted from the 615, that leaves 560 for department II's capitalists to actually consume. So, once every transfer — the ones already made and the ones still pending — has gone through, we get, in capital value:

For things to go normally, department II's accumulation has to move faster than department I's. Otherwise, the part of department I's variable capital plus surplus value that has to be exchanged for department II's goods would grow faster than department II's constant capital, which is what that part has to be exchanged against.

If reproduction continues on this basis, with everything else staying the same, then at the close of the following year we get:

With the split of surplus value staying the same: department I first has to spend, as revenue, 1,210 in variable capital plus half its surplus value — 605 — 1,815 together. That consumption fund is again 55 more than department II's constant capital. The 55 has to be taken out of department II's 880 in surplus value, leaving 825. Turning that 55 into department II's constant capital also means a further deduction from its surplus value, for the matching variable capital — 27½ — leaving 797½ for department II to consume.

Now 605 in surplus value has to be capitalized in department I: 484 of it as constant capital, 121 as variable capital. That 121 has to be taken from department II's surplus value, which now stands at 797½, leaving 676½. So department II turns a further 121 into constant capital, and needs a further 60½ in variable capital for it; this too comes out of the 676½, leaving 616 for consumption.

We then have, in capital:

And at the end of the year, in product:

Repeating the same calculation, and rounding off the fractions, at the close of the following year we get a product of:

And at the close of the year after that:

Over five years of reproduction on an expanded scale, the combined capital of departments I and II has risen from 5,500 in constant capital plus 1,750 in variable capital — 7,250 together — to 8,784 in constant capital plus 2,782 in variable capital — 11,566 together. That's a ratio of 100 to 160. Total surplus value started at 1,750; it now stands at 2,782. The surplus value actually consumed started at 500 for department I and 600 for department II — 1,100 together; in the last year it was 732 for department I and 745 for department II — 1,477 together. So it has grown in a ratio of 100 to 134.

Take the year's whole product now: 9,000, all of it sitting as commodity capital in the hands of the industrial capitalist class, in a form where the general average ratio of variable to constant capital is 1 to 5.

That ratio assumes some things are already true: capitalist production, and with it the productive power of social labour, is already significantly developed; the scale of production has already been significantly expanded before this; and, finally, all the conditions are in place that produce a relative surplus population within the working class — part of it kept in reserve, without work.

Rounding off the fractions, the year's product then divides up as follows:

Now suppose the capitalist class in department I consumes half its surplus value — 500 — and saves the other half to accumulate. Then 1,500 from department I (1,000 in variable capital plus 500 in surplus value) would need to be exchanged for 1,500 of department II's constant capital.

But department II's constant capital comes only to 1,430, so an extra 70 has to be added out of surplus value. Deducted from department II's 285 in surplus value, that leaves 215. So we get:

Department I: 5,000 in constant capital, plus 500 in surplus value still to be turned into capital, plus 1,500 — variable capital plus surplus value — in the capitalists' and workers' consumption fund.

Department II: 1,430 in constant capital, plus 70 in surplus value still to be turned into capital, plus 285 in variable capital, plus 215 in surplus value.

Since this 70 of department II's surplus value is added directly onto its constant capital, setting that extra constant capital to work requires additional variable capital too. At the ratio of 1 to 5, that means 70 divided by 5 — 14. So a further 14 comes out of the 215 left in department II's surplus value, leaving 201. We then have:

Exchanging 1,500 of department I's variable capital plus half its surplus value for 1,500 of department II's constant capital is, on its own, just simple reproduction — settled already, as far as that goes.

Still, a few peculiarities need pointing out here. They come from the fact that under accumulating reproduction, department I's variable capital plus half its surplus value is not replaced by department II's constant capital alone, but by that constant capital plus part of department II's surplus value.

It's easy to see why, once department I is accumulating, its variable capital plus surplus value has to be bigger than department II's constant capital — not equal to it, the way simple reproduction requires. There are two reasons. First, department I keeps part of its own surplus product for its own productive capital, and turns five-sixths of that part into constant capital; that five-sixths can't also be replaced, at the same time, by department II's consumption goods. Second, department I has to supply the material for the extra constant capital that accumulation requires inside department II — just as department II has to supply the material for the variable capital that sets in motion the part of department I's own surplus product that department I is using as extra constant capital.

Here it matters what variable capital actually is: real variable capital consists of labour power, and so does the additional variable capital. It is not the capitalist in department I who buys up or stockpiles provisions from department II in advance, for the extra labour power he intends to take on — a slave-holder had to do that. It is the workers themselves who deal with department II.

That doesn't stop the capitalist from seeing those purchases differently, though. From his standpoint, the means of consumption that additional labour power will buy are simply the means of producing and maintaining whatever extra labour power he may take on — in other words, the natural form his variable capital takes.

His own actual next task — here, department I's — is only to hoard the new money capital needed to buy that additional labour power. Only once he has actually taken the labour power on does this money become a means of buying department II's goods for it, and only then must those means of consumption already be there waiting.

By the way: the capitalist gentleman, like his press, is often unhappy with how labour power spends its money — and with the goods from department II it spends that money on. On occasions like this he turns philosopher, culture-talker, and philanthropist all at once. Mr Drummond, for instance — a British diplomat in Washington, secretary of the legation there — reports that The Nation, a newspaper, had carried an interesting article in October 1879, which said, among other things:

'Workers have not kept pace, in matters of culture, with the progress of invention. Masses of things have become available to them that they don't know how to use, and so create no market for.' {Naturally every capitalist wants the worker to buy his goods.} 'There is no reason why the worker shouldn't want as many comforts as the clergyman, lawyer, or doctor who earns the same amount he does.' {That sort of lawyer, clergyman and doctor does indeed have to stop at wishing for plenty of comforts!} 'But he doesn't. The question remains how he is to be raised as a consumer through a rational and healthy procedure — no easy question, since his whole ambition goes no further than shortening his working hours, and the demagogue eggs him on to that far more than to raising his condition by improving his intellectual and moral capacities.'

Long working hours seem to be the secret of this rational and healthy procedure — the one that's supposed to raise the worker's condition by improving his intellectual and moral capacities, and turn him into a rational consumer. To become a rational consumer of the capitalists' goods, he must first — but the demagogue stops him! — let his own capitalist consume his own labour power irrationally and unhealthily.

What the capitalist actually means by rational consumption shows itself wherever he condescends to step directly into his workers' spending — in the truck system, paying wages in goods redeemable only at the company's own store, and in supplying workers' housing, so that the same capitalist is also their landlord: just one branch of the business among many.

That same Drummond — the one whose fine feelings wax enthusiastic over these capitalist attempts to uplift the working class — reports, elsewhere in the same account, on the cotton mills at Lowell and Lawrence. The boarding houses where the mill girls eat and sleep belong to the joint-stock company that owns the factory; the women running these houses are employed by that same company, which lays down rules of conduct for them; no girl is allowed to come home after ten at night.

But here is the pearl of it: the company runs its own special police, patrolling the area to stop this house rule being broken. After ten in the evening, no girl is let out or let back in. No girl may lodge anywhere except on land the company owns, where every house brings it about $10 a week in rent. And now, in full glory, here is the rational consumer:

'Since the ever-present piano turns up in many of the best lodging houses for working women, music, singing, and dancing play a considerable part — at least for those who, after ten hours steadily at the loom, need more variety from the monotony than they need real rest.'

But the chief secret of how to turn a worker into a rational consumer is still to come. Mr Drummond visits the cutlery factory at Turner's Falls, on the Connecticut River, and Mr Oakman, the treasurer of the joint-stock company, after telling him that American table-knives in particular beat the English on quality, goes on:

'We shall beat England on price too. We're already ahead of them on quality — that's acknowledged. But we need lower prices, and we'll get them as soon as we've got our steel cheaper and beaten down our labour!'

Cutting wages and lengthening working hours — that is the whole substance of this rational and healthy procedure, meant to raise the worker to the dignity of a rational consumer, so that he creates a market for the mass of things that culture and the progress of invention have put within his reach.

Just as department I has to supply department II's extra constant capital out of its own surplus product, so department II, in the same way, supplies department I's extra variable capital. Where variable capital is concerned, department II accumulates for both departments — itself included — simply by reproducing a bigger share of everything it makes, its surplus product especially, in the form of necessary means of consumption.

When production runs on a growing capital basis, department I's variable capital plus its surplus value must equal: department II's constant capital, plus whatever part of department II's surplus product gets folded back into capital, plus the extra constant capital department II needs to expand its production. There is a floor under that last piece — a minimum expansion — and without at least that much, genuine accumulation, meaning the actual extension of production in department I itself, cannot happen.

Let's go back to the case just considered. Its peculiarity: department II's constant capital is smaller than department I's wages plus half its surplus value — smaller, that is, than the part of department I's product spent as revenue on means of consumption. So turning over department I's full 1,500 requires realizing part of department II's surplus product as well — 70 worth. As for the remaining 1,430 of department II's constant capital: other things being equal, it simply has to be replaced out of department I's wages and surplus value, at the same value, for simple reproduction to happen in department II — and that settles it, nothing more to say.

The extra 70 is different. Follow the same trade from both sides and it means two different things at once. For department I, it is just the exchange of revenue for means of consumption — an exchange aimed only at consumption. For department II here, it is not — as it would be under simple reproduction — merely turning constant capital back from the form of commodity capital into its own natural form. It is instead the actual process of accumulation itself: part of II's surplus product converted from the form of means of consumption into that of constant capital.

Suppose department I uses £70 in money — its money reserve for turning over surplus value — to buy that 70 of department II's surplus product. And suppose department II does not use the money to buy 70 of department I's surplus product back, but instead accumulates the £70 as money capital. That money capital would still be the expression of extra product — precisely department II's own surplus product, a fractional part of it — even though not of a product that goes back into production. But then this accumulation of money on department II's side would, at the very same time, be the expression of an unsaleable 70 of department I's surplus sitting as means of production. There would then be relative overproduction in department I, matching this very failure of department II to expand its own reproduction.

But apart from this: for as long as the £70 in money that came from department I has not yet returned to department I — because department II has not yet bought, or has only partly bought, that 70 of I's surplus product back with it — the £70 counts, wholly or partly, as additional virtual money capital sitting in department II's hands. That is true of every exchange between the two departments, right up until each side's goods have replaced the other's and sent the money back to where it started. Under normal conditions, though, the money holds this role only briefly.

In the credit system, where any bit of money set free even for a moment is supposed to spring straight into action as additional money capital, this only-temporarily-free money capital can get tied up — used, say, for new enterprises within department I — when it ought instead to be setting in motion surplus product that is still sitting stuck, unsold, in other enterprises.

There is also this to note: annexing that 70 to department II's constant capital at the same time requires department II's variable capital to expand too, by 14. This presupposes — just as the direct folding of surplus product into constant capital does in department I — that reproduction in department II is already under way with a tendency toward further capitalization, and so already includes an expansion of the part of the surplus product made up of necessary means of subsistence.

Take the 9,000 product from the second example: as we already saw, it has to be divided up in the following way for reproduction to happen — provided 500 of department I's surplus value is to be capitalized. Here we consider only the goods themselves, and leave money circulation aside.

Department I: 5,000 in constant capital, plus 500 in surplus value still to be capitalized, plus 1,500 — variable capital plus surplus value — as the consumption fund. That's 7,000 in commodities.

Department II: 1,500 in constant capital, plus 299 in variable capital, plus 201 in surplus value. That's 2,000 in commodities — 9,000 in commodity product altogether.

The capitalizing now proceeds as follows:

In department I, the 500 in surplus value being capitalized splits five-sixths to one-sixth: 417 becomes constant capital, 83 becomes variable capital. That 83 draws an equal amount out of department II's surplus value, which buys elements of constant capital and gets added to department II's constant capital. An increase of 83 in department II's constant capital calls for an increase of one-fifth of that — 17 — in department II's variable capital. We then have, after the exchange:

Department I's capital now functions at 6,500 where it was 6,000 — a rise of one-twelfth. Department II's has grown from 1,715 to 1,899 — just under one-ninth.

Reproduction on this basis in the second year yields, at year's end, in capital:

And at the end of the third year, in product:

If department I again accumulates half its surplus value here, as before, then department I's wages plus half its surplus value comes to 1,173 in variable capital plus 587 — half the surplus — making 1,760: bigger than the whole of department II's constant capital, 1,715, by 45. That 45 must, again, be balanced out by transferring an equal amount of means of production onto department II's constant capital. Department II's constant capital thus grows by 45, which calls for an increase of one-fifth of 45 — 9 — in its variable capital.

The capitalized 587 of department I's surplus value then splits five-sixths to one-sixth: 489 becomes constant capital, 98 becomes variable capital. That 98 calls for a fresh addition of 98 to department II's constant capital as well, and this in turn calls for an increase of one-fifth of 98 — 20 — in department II's variable capital. We now have:

Over three years of growing reproduction, department I's total capital has grown from 6,000 to 7,629, department II's from 1,715 to 2,229, and the total social capital from 7,715 to 9,858.

So the exchange between I(v+s) and IIc can go several different ways.

Under simple reproduction, the two sides must be equal and must replace each other — otherwise, as we've already seen, simple reproduction can't proceed without disruption.

Under accumulation, the first thing to consider is the rate of accumulation itself. In the examples used so far, department I's rate of accumulation was always half its surplus-value, held constant from year to year. The only thing that changed was how that accumulated capital splits between new variable capital and new constant capital. That gives three cases:

Case 1: I(v+½s) equals IIc — a sum that's smaller than the whole of I(v+s). That gap is what always has to hold, not the exact match: if IIc were not smaller than I(v+s), department I would not be accumulating at all.

Case 2: I(v+½s) is bigger than IIc. Here the shortfall gets covered by adding a matching part of IIs to IIc, until the two together equal I(v+½s). For department II, this exchange is no longer simple replacement of its constant capital — it's already accumulation: department II is growing its constant capital by the part of its surplus product it trades for department I's means of production. And that growth comes bundled with more: department II also enlarges its variable capital out of that same surplus product.

Case 3: I(v+½s) is smaller than IIc. Here the exchange leaves department II's constant capital not fully replaced, so department II has to make up the shortfall by buying more from department I. That purchase doesn't call for any further accumulation of variable capital in department II — it only brings department II's constant capital up to its full size, nothing more.

But look at the other side of the same exchange: for the section of department I's capitalists who are simply piling up additional money capital, this sale has already done part of that kind of accumulating for them.

The condition for simple reproduction — that I(v+s) exactly equal IIc — doesn't fit capitalist production, and that's true for two separate reasons. First: this incompatibility doesn't rule out something different that's also real — within the roughly ten-to-eleven-year industrial cycle, some years actually produce less than the year before, so little that not even simple reproduction happens relative to the previous year. Second: given ordinary yearly population growth, simple reproduction would mean an ever-larger number of unproductive retainers sharing in the 1,500 that stands for total surplus-value. Real accumulation of capital — genuine capitalist production — would be impossible on those terms. So the fact that capitalist accumulation happens at all rules out IIc equalling I(v+s).

Even so, under capitalist accumulation itself, something else could still happen: through the accumulation carried out over an earlier run of production periods, IIc could end up not just equal to I(v+s) but actually bigger. That would mean overproduction in department II — fixable only by a major crash, one that would shift capital from department II over to department I.

None of this changes the relation between I(v+s) and IIc if part of department II's constant capital is reproduced within department II itself — in agriculture, say, by sowing home-grown seed. That self-reproduced part of IIc plays no role at all in the exchange between department I and department II, no more than Ic does. Nor does it change anything if part of what department II produces can itself serve as means of production in department I. That part is covered by a part of the means of production department I supplies — and both these matched parts have to be deducted from both sides at the outset, if we want to examine the exchange between the two great departments of social production, the producers of means of production and the producers of means of consumption, in its pure, unclouded form.

So under capitalist production, I(v+s) can never simply equal IIc — the two sides can't balance each other in this exchange. But let Is/x stand for the part of Is that department I's capitalists spend as revenue rather than accumulate: then I(v+s/x) can equal, exceed, or fall short of IIc — all three stay open. What can never happen: I(v+s/x) reaching all the way up to II(c+s). It always falls short — short by exactly the part of IIs that department II's capitalists have to consume themselves, no matter what.

One thing to flag: this whole account of accumulation doesn't represent the value of constant capital exactly, in so far as that value is a piece of the commodity capital it helps produce. The fixed part of newly accumulated constant capital only enters commodity capital gradually, in instalments — differently depending on what kind of fixed element it is. So wherever raw material and semi-finished goods go into commodity production in bulk, that commodity capital mostly consists of replacements for the circulating constant capital and the variable capital instead.

This way of proceeding still works because of how the circulating components turn over: it assumes that within the year, the circulating part, together with the share of fixed capital's value handed on to it, turns over often enough that the total of commodities supplied equals the value of the whole capital that goes into that year's production.

But where, as in running machinery, only ancillary materials enter and no raw material at all, the labour element — variable capital — has to show up again as the larger component of the commodity capital instead. And there's a further contrast: the rate of profit calculates surplus-value on the whole capital, regardless of whether the fixed components hand over a lot of value to the product in a given period or only a little. But for the value of any commodity capital actually produced, the fixed part of constant capital only counts in so far as it actually gives up value to the product through average wear and tear.

Department II's original source of money is the wages-plus-surplus of gold production, which sits inside department I, exchanged for part of IIc. That money reaches department II only in part: to the extent that the gold producers store up surplus-value, or convert it into department I's own means of production — that is, expand their own output — that much of their wages-plus-surplus does not go into department II.

On the other hand, once the gold producers' own accumulation of money eventually leads to expanded reproduction, the part of gold production's surplus-value that isn't spent as revenue — meant instead for the gold producers' additional variable capital — does go into department II. There it either calls for fresh hoard formation, or supplies new means to buy from department I without selling straight back to it.

From the money that comes from this I(v+s) of gold production, subtract whatever gold certain branches of department II need as raw material and the like — in short, as a replacement element of their own constant capital.

In the exchange between department I and department II, an element counts as provisional hoard formation — building up for the sake of future expanded reproduction — only in these cases: in department I, when part of Is is sold to department II one-sidedly, with no purchase back the other way, and serves there as additional constant capital for department II; in department II, when department I buys one-sidedly for additional variable capital; and further, whenever part of the surplus-value department I spends as revenue isn't covered by department II, so that part of IIs gets bought instead and turned into money that way.

If I(v+s/x) turns out bigger than IIc, then IIc doesn't need any separate top-up in goods from department I to replace what department I has already drawn out of IIs for its own simple reproduction. That raises a further question: how far can hoard formation happen within the exchange of department II's own capitalists among themselves — an exchange that can only consist of trading IIs back and forth?

We already know that within department II, direct accumulation happens only when part of IIs is converted straight into variable capital — just as, within department I, part of Is is converted straight into constant capital. Given the different stages of accumulation across department II's various lines of business, and among the individual capitalists within each line, the matter works out, changed only where it must, exactly as it did for department I: some capitalists are still at the stage of hoard formation, selling without buying; others, having reached the point of actually expanding reproduction, buy without selling.

The additional variable money capital is certainly laid out at first on additional labour-power. But that labour-power buys means of subsistence from the hoard-forming owners of the extra means of consumption that go into workers' consumption. And from those owners, in proportion to how much they're hoarding, the money does not return to where it started — they simply store it up.