In this chapter and the next one, we look at how the time it takes capital to turn over affects how much value that capital can create.

Take a batch of goods that is the product of one working period — say, nine weeks. For now, set aside two things: the extra value added by the ordinary wear of fixed capital, and the surplus-value added during production. With those set aside, the value of this batch of goods equals the value of the circulating capital advanced to produce it — that is, the wages paid and the raw and auxiliary materials used up. Say that value is £900, so the weekly outlay is £100. The periodic production time, which here is the same as the working period, is therefore 9 weeks. It makes no difference whether we think of this as one working period producing a single continuous product, or as a continuous working period producing a series of separate items — what matters is only that the quantity of goods taken to market at one time costs 9 weeks of labour to produce. Say the circulation time lasts 3 weeks. The whole turnover period then lasts 12 weeks. After 9 weeks the advanced productive capital has been turned into commodity capital, but it now spends three weeks sitting in circulation. So a new round of production cannot begin again until the start of the 13th week, and production would stand still for three weeks — one quarter of the whole turnover period. Again it makes no difference whether we assume it simply takes that long, on average, to sell the goods, or that the time is set by the distance to market, or by the payment terms on the goods sold. Every three months production would stand still for three weeks; over the year that is 4 × 3 = 12 weeks = 3 months = one quarter of the year's turnover time.

So if production is to run continuously, week after week, on the same scale, only two things are possible.

One option — not a neutral one: shrink the scale of production, so that the £900 is enough to keep work going both during the working period and during the circulation time of the first turnover. Then, in week 10, a second working period — and so a second turnover period — begins before the first turnover period has even finished, since the turnover period is twelve weeks long but the working period only nine. Spread over 12 weeks, £900 works out to £75 a week (and note this route ties up relatively more of the capital as idle money, not less). This has consequences. First, shrinking the scale of the business this way means changing the size of the fixed capital too — the whole business set-up has to be scaled down. Second, it is not even clear this shrinking is always possible: for any given branch of production, there is a normal minimum amount of capital needed, and a business below that minimum cannot compete. This minimum itself keeps growing as capitalist production develops, so it is not fixed. But between whatever the normal minimum is at a given time and the ever-expanding normal maximum, there are many intermediate steps — a middle range that allows for very different amounts of capital. Within that middle range, shrinking the scale is possible; its floor is simply whatever the normal minimum happens to be at the time. Separately: when production is checked, markets are glutted, or raw materials become dear, the normal outlay of circulating capital — with the fixed capital unchanged — gets restricted by cutting working time, for instance by working only half-days. In the same way, in times of prosperity, the circulating capital gets abnormally extended on the same fixed-capital base, partly by lengthening the working day and partly by intensifying it. In businesses set up from the outset to handle such swings, people manage partly by these same means and partly by using more workers at once together with reserve fixed capital — reserve locomotives on the railways, for example. But such abnormal swings are set aside here, since we are assuming normal conditions.

So to make production continuous, the outlay of this same circulating capital gets spread over a longer stretch of time — 12 weeks instead of 9. At any given moment, then, a reduced productive capital is at work; the liquid part of the productive capital is cut from £100 to £75, a reduction of one quarter. The total amount by which the productive capital at work during the 9-week working period is reduced comes to 9 × £25 = £225, or one quarter of £900. And the ratio of the circulation time to the whole turnover period is likewise 3/12 = 1/4. So here is what follows: if production is not to be interrupted during the circulation time of the productive capital that has turned into commodity capital — if it is instead to continue, week after week, without a break — and if no separate circulating capital is provided for this, then the only way to achieve it is by cutting back the scale of production, that is, by shrinking the liquid part of the productive capital at work. The liquid part released from the original advance — a portion of the same £900, not newly earned money — that becomes available for use during the circulation time stands to the whole advanced liquid capital in the same ratio as the circulation time stands to the turnover period. This holds, as already noted, only for branches of production where the labour process is carried out week after week on the same scale — not for branches like agriculture, where different working periods require different sums of capital to be laid out.

Now assume the opposite: that the way the business is set up rules out shrinking the scale of production, and so also rules out shrinking the liquid capital that must be advanced each week. Then continuous production can only be achieved by means of additional liquid capital — in the example above, £300. Over the 12-week turnover period, £1,200 is advanced in stages, of which £300 is the fourth part, just as 3 weeks is the fourth part of 12. After the 9-week working period, the capital-value of £900 has been converted from the form of productive capital into the form of commodity capital. Its working period is finished, but it cannot be renewed with the same capital. During the three weeks it spends in the sphere of circulation, functioning as commodity capital, it is — as far as the production process is concerned — exactly as if it did not exist at all. We are leaving all credit relations out of account here, and therefore assuming that the capitalist operates only with his own capital. But while the capital advanced for the first working period spends three weeks in the circulation process once production is complete, an additional capital of £300 is laid out and put to work, so that the continuity of production is not interrupted.

Now the following should be noted here:

First: the working period of the £900 first advanced is finished after 9 weeks, and it does not flow back before 3 more weeks have passed — that is, not until the start of the 13th week. But a new working period opens again immediately, using the additional £300. This is exactly what establishes the continuity of production.

Second: the roles of the original £900 and of the £300 laid out as extra capital at the end of the first 9-week working period — the £300 that opens the second working period the moment the first ends, without a break — are cleanly separated during the first turnover period, or at least can be. But over the course of the second turnover period, the two capitals cross over each other.

Let's picture the matter in concrete terms:

First turnover period, 12 weeks. First working period, 9 weeks; the turnover of the capital advanced here is completed at the start of the 13th week. During the last 3 weeks, the additional capital of £300 is at work, and it opens the second working period of 9 weeks.

Second turnover period. At the start of the 13th week, £900 has flowed back and is able to begin a new turnover. But the second working period was already opened in week 10 by the additional £300; by the start of the 13th week, a third of that working period is already finished, £300 having been converted from productive capital into product. Since only 6 more weeks are needed to finish the second working period, only two-thirds of the returned £900 — that is, only £600 — can go into the production process of the second working period. £300 of the original £900 is set free, to play the same role that the additional £300 played during the first working period. By the end of week 6 of the second turnover period, the second working period is complete. The £900 laid out in it flows back after 3 weeks — that is, at the end of week 9 of the second, twelve-week turnover period. During the 3 weeks of its circulation time, the £300 that was set free steps in. This is what starts the third working period of a £900 capital, in week 7 of the second turnover period, or week 19 of the year.

Third turnover period. At the end of week 9 of the second turnover period, £900 flows back again. But the third working period already began in week 7 of the previous turnover period, and 6 weeks of it have already gone by. So it only lasts 3 more weeks. Of the £900 that has just flowed back, only £300 goes into the production process. The fourth working period fills the remaining 9 weeks of this turnover period, so that the fourth turnover period and the fifth working period both begin together, in week 37 of the year.

To make the calculation simple, suppose: a production period of 5 weeks, a circulation time of 5 weeks, so a turnover period of 10 weeks — counting the year as 50 weeks, with £100 advanced per week. The production period thus needs a fluid capital of £500, and the circulation time needs a further £500 of additional capital. Production periods and turnover periods then run as follows:

If the circulation time is zero, the turnover period is simply the production period, so the number of turnovers in a year equals the number of production periods in a year. With a five-week production period, that is 50 divided by 5, or 10, and the value of the capital turned over would be 500 times 10, that is, £5,000. In the table above, where the circulation time is taken as 5 weeks, the same £5,000 worth of goods gets produced each year — but one-tenth of it, £500, is always sitting in the form of commodity-capital, and only flows back after 5 weeks. So by the end of the year, the output of the tenth production period (weeks 46 to 50) has completed only half its turnover time, because its circulation time falls into the first 5 weeks of the next year.

Let's take a third example: a production period of 6 weeks, a circulation time of 3 weeks, and £100 advanced each week in the production process.

First production period: weeks 1 to 6. By the end of week 6 there is a commodity-capital of £600, which flows back at the end of week 9.

Second production period: weeks 7 to 12. During weeks 7 to 9, £300 of additional capital is advanced. At the end of week 9, £600 flows back. Of that, £300 is advanced for weeks 10 to 12; so by the end of week 12, £300 is free in liquid form, £600 sits in commodity-capital, and it flows back at the end of week 15.

Third production period: weeks 13 to 18. In weeks 13 to 15, the £300 just freed is advanced; then £600 flows back, of which £300 is advanced for weeks 16 to 18. By the end of week 18, £300 is free as money; £600 sits in commodity-capital, flowing back at the end of week 21. (A fuller account of this case follows further on, as Case II.)

So over 9 production periods (54 weeks), £600 times 9, or £5,400 worth of goods, gets produced. By the end of the ninth production period, the capitalist holds £300 in money and £600 in goods that have not yet finished their circulation time.

Comparing the three examples, we find, first, that only in the second one do capital I (£500) and additional capital II (also £500) take turns cleanly, moving as two separate blocks. That happens only because this example makes the quite exceptional assumption that the production period and the circulation time are two equal halves of the turnover period. In every other case — whatever the imbalance between the two parts of the turnover period — the movements of the two capitals cross over each other, as in examples I and III, starting from the second turnover period on. Additional capital II then joins with part of capital I to form the capital at work in the second turnover period, while the rest of capital I is freed up for the job capital II was originally advanced for. The capital active during the commodity-capital's circulation time is, at this point, not literally the same capital II originally advanced for that purpose — but it equals it in value and makes up the same fraction of the total capital advanced.

Second: the capital that has been at work during the production period lies idle during the circulation time. In the second example, the capital works for the 5-week production period and lies idle for the 5-week circulation time. Added up over the year, that comes to half a year that capital I spends idle. Additional capital II steps in to cover that idle stretch — and so it, too, in this case, lies idle for half a year. But the additional capital needed to keep production continuous through the circulation time is not fixed by the total amount of circulation time added up across the year. It is fixed only by the ratio of the circulation time to the turnover period. (This assumes, of course, that every turnover happens under the same conditions.) That is why example II needs £500 of additional capital, not £2,500. The reason is simple: the additional capital enters into the turnover just as the original capital does, and so, exactly like the original capital, it replaces its own bulk through the number of times it turns over.

Third: whether the production time is longer than the labour time makes no difference to what we have been looking at here. That does lengthen the turnover periods overall, but this lengthened turnover calls for no additional capital for the labour process itself. The additional capital exists for one purpose only: to fill the gaps torn in the labour process by the circulation time. Its job is to shield production from disruptions that come from the circulation time — disruptions that come from production's own conditions have to be dealt with in some other way, one we are not considering here. There are, on the other hand, businesses that work only in bursts, on order, so that gaps can open up between one production period and the next. In such businesses, the need for additional capital falls away, to that extent. On the other hand, in most cases of seasonal work there is also some limit on how long the return can take. The same work cannot be started again next year with the same capital if that capital's circulation time has not run its course by then. But the circulation time can also be shorter than the gap between one production period and the next. In that case the capital lies idle, unless it is put to some other use in the meantime.

The capital advanced for one production period — the £600 in example III, say — is laid out partly in raw and auxiliary materials, as a productive stock for the production period, that is, as constant circulating capital, and partly as variable circulating capital, in payment for labour itself. The part laid out as constant circulating capital need not sit as productive stock for the same stretch of time throughout — the raw material, for instance, might not lie there for the whole production period, and coal might only be brought in every two weeks. Still, since credit is ruled out here, whatever part of this capital is not available in the form of productive stock has to remain available in money form, ready to be converted into productive stock as needed. This changes nothing about the size of the constant circulating capital-value advanced for the 6 weeks. Wages, though — leaving aside the money reserve for unforeseen expenses, the actual reserve fund for smoothing out disruptions — are paid out over shorter periods.

Wages are usually paid weekly. So unless the capitalist forces the worker to extend credit on their own labour, the capital needed for wages has to be on hand in money form. This means that when the capital flows back, part of it has to be kept in money form to pay for labour, while the rest can be converted into productive stock.

Additional capital II splits up into the same parts as the original capital. But what sets it apart from capital I is this: leaving credit relations aside, in order to be ready for its own production period, it has to be advanced already during the whole of capital I's first production period — a period it plays no part in. During that time, at least part of it can already be converted into constant circulating capital, advanced for the whole turnover period. How far it takes that form, or how far it stays as additional money-capital until the moment conversion becomes necessary, depends partly on the particular production conditions of the branch of business, partly on local circumstances, partly on price swings in raw materials, and so on. Looking at the total social capital, some more or less sizeable part of this additional capital will always sit in the state of money-capital for a longer stretch of time. The part of capital II earmarked for wages, though, is different: it only gets converted into labour-power gradually, as each shorter production period runs its course and gets paid. So this part of capital II exists in money-capital form for the entire length of the production period, until it is converted into labour-power and enters into the function of productive capital.

So the arrival of this additional capital — needed to turn capital I's circulation time into production time — does more than just increase the size of the capital advanced and the length of time for which the total capital necessarily has to be advanced. It specifically increases the part of the advanced capital that exists as a money reserve — that is, the part sitting in the state of money-capital, in the form of potential money-capital.

The same thing happens — both for the part advanced as productive stock and for the part advanced as a money reserve — when the split that the circulation time forces on the capital, into capital for the first production period and replacement capital for the circulation time, comes about not by enlarging the capital laid out, but by shrinking the scale of production instead. Relative to the scale of production, the growth in the capital locked up in money form is, if anything, even greater on this route.

What this whole splitting of the capital into an originally productive part and an additional part achieves is simply the unbroken succession of production periods — the constant functioning, as productive capital, of an equally sized part of the capital advanced.

Let's look at example II. The capital constantly at work in the production process is £500. Since the production period is 5 weeks, it works ten times over during the 50 weeks we are counting as a year. So the output, setting surplus-value aside, comes to 10 times 500, or £5,000. From the standpoint of the capital directly and continuously at work in the production process — a capital-value of £500 — the circulation time appears to have vanished completely. The turnover period appears to coincide with the production period; the circulation time appears to be set at zero.

But suppose instead that this £500 capital were regularly held back in its productive activity by the 5-week circulation time, so that it could only become fit for production again once the whole 10-week turnover period was over. Then in the 50 weeks of the year we would have 5 turnovers of ten weeks each — made up of 5 production periods of 5 weeks, so 25 production-weeks in all, with a total output of 5 times 500, or £2,500; and 5 circulation times of 5 weeks, so a total circulation time likewise of 25 weeks. If we now say that the £500 capital turned over five times in the year, it is plain and clear that for half of every turnover period, this £500 capital did not function as productive capital at all — and that, added up, it functioned for only half the year, and not at all during the other half.

In our example, the £500 replacement capital steps in for the duration of these five circulation times, and that raises the turnover from £2,500 to £5,000. But the capital advanced is now £1,000 instead of £500. £5,000 divided by £1,000 is 5. So five turnovers instead of ten — and this is in fact how it gets calculated. But once people start saying that the £1,000 capital "turned over five times in the year," the memory of the circulation time vanishes from the capitalists' hollow skulls, and a confused idea takes hold — as if this capital had been constantly at work in the production process throughout those five successive turnovers. Yet if we say this £1,000 capital turned over five times, that figure includes both circulation time and production time. If £1,000 really had been continuously at work in the production process, the output, on our assumptions, would have to be £10,000, not £5,000. And to keep £1,000 continuously at work in production, a full £2,000 would have to be advanced in the first place. The economists — among whom nothing clear about the mechanism of turnover is to be found at all — constantly overlook this key point: that only part of industrial capital can ever actually be engaged in the production process at any moment, if production is to proceed without interruption. While one part is in its production period, some other part always has to be in its circulation period. Put differently, one part can function as productive capital only on condition that another part, in the form of commodity-capital or money-capital, stays withdrawn from actual production. And by overlooking this, they overlook the significance and role of money-capital altogether.

Now we need to look at what difference shows up in turnover depending on whether the two parts of the turnover period — the working period and the circulation period — are equal to each other, or the working period is longer or shorter than the circulation period. And further, how this affects the amount of capital tied up in the form of money-capital.

We assume that in every case the capital to be advanced each week is £100, and the turnover period is 9 weeks — so the capital to be advanced for each turnover period is £900.

This case, although in reality only an accidental exception, has to serve as our starting point, because here the relations show themselves in the simplest and most tangible way.

The two capitals — Capital I, advanced for the first working period, and additional Capital II, which does its work during Capital I's circulation period — take turns in their movements without ever crossing each other. So, apart from the very first period, each of the two capitals is only ever advanced for its own turnover period. Let the turnover period be, as in the examples that follow, 9 weeks, so the working period and the circulation period are each 4 1/2 weeks. Then we get the following scheme for the year:

Within the 51 weeks that we are counting here as the year, Capital I has completed six full working periods, producing, for 6 ×

£450 = £2,700; and Capital II, in five full working periods, for 5 ×

£450 = £2,250 worth of goods. On top of that, Capital II produced a further £150 worth in the last 1 1/2 weeks of the year (from the middle of week 50 to the end of week 51) — total product over the 51 weeks: £5,100. If we look only at the direct production of surplus-value, which happens only during the working period, the total capital of £900 would then have turned over five whole times plus two-thirds of a sixth period (5 2/3 × 900 = £5,100). But if we look at the actual turnover, Capital I has turned over five whole times plus two-thirds of its sixth turnover period, since at the end of week 51 it still has 3 weeks of its sixth turnover period left to run; 5 2/3 × 450 = £2,550. And Capital II has turned over five whole times plus one-sixth of its sixth turnover period, since it has completed only 1 1/2 weeks of that sixth period, leaving 7 1/2 weeks of it falling into the next year; 5 1/6 × 450 = £2,325. The actual total turnover is £4,875.

Let us consider Capital I and Capital II as two capitals entirely independent of each other. In their movements they really are entirely independent; these movements only complement each other because their working periods and circulation periods directly relieve one another in turn. They can be treated as two wholly independent capitals, belonging to two different capitalists.

Capital I has completed five whole turnover periods, plus two-thirds of its sixth. At the end of the year it is sitting there in the form of commodity-capital, and it still needs 3 more weeks before it can be normally realized. During that time it cannot re-enter the production process. It is functioning as commodity-capital: it is circulating. Of its last, sixth turnover period, it has covered only two-thirds. This is put by saying it has turned over only 2/3 of a time — that only 2/3 of its total value has completed a full turnover. We say: £450 completes its turnover in 9 weeks, so £300 completes a turnover in 6 weeks. But this way of putting it glosses over the organic relation between the two specifically different parts that make up turnover time. The exact meaning of saying that the advanced capital of £450 has made 5 2/3 turnovers is only this: it has completed five turnovers in full, and of the sixth, only two-thirds. By contrast, the statement that the turned-over capital equals 5 2/3 times the advanced capital — in the case above, 5 2/3 × £450 = £2,550 — is correct in a different sense: it captures that if this £450 were not backed up by another £450, part of it would indeed have to be sitting in the production process while another part sat in the circulation process at the same time. If turnover time is to be expressed as a mass of turned-over capital, it can only ever be expressed as a mass of value that already exists — in fact, of finished product. The fact that the advanced capital is not in a state where it can open a new round of production by itself shows up in this: only part of it is fit to produce, or else, to keep production continuous, the capital would have to be split — one part permanently sitting in the production period, the other permanently in the circulation period, in whatever ratio those two periods stand to each other. It is the very same law that fixes the mass of productive capital that is constantly at work, by the ratio of circulation time to turnover time.

Of Capital II, by the end of week 51 — which we are treating here as the close of the year — £150 has been advanced into producing unfinished product. A further part is sitting in the form of fluid constant capital — raw materials and the like — that is, in a form in which it can function as productive capital in the production process. But a third part is sitting in money form: at least the amount of wages for the rest of the working period (3 weeks), which, however, is only paid at the end of each week. Now even though this part of the capital, at the start of the new year — a new turnover cycle — is not in the form of productive capital but of money-capital, and in that form cannot enter the production process, all the same, when the new turnover opens, fluid variable capital — that is, living labour-power — is already at work in the production process. This happens because labour-power is bought and used up at the start of the working period, say by the week, but is only paid for at the end of the week. Here money functions as means of payment. So on one side the money is still sitting in the capitalist's hand, while on the other side the labour-power — the commodity it is to be exchanged for — is already at work in the production process. The very same capital-value thus appears here twice at once.

If we look only at the working periods, Capital I has produced 6 × £450 = £2,700. Capital II has produced 5 1/3 × £450 = £2,400. Together, that comes to 5 2/3 × £900 = £5,100.

So the total advanced capital of £900 has functioned as productive capital 5 2/3 times over the year. Whether it is always £450 in the production process and £450 in the circulation process, alternating, or whether it is the full £900 in the production process for 4 1/2 weeks and then in the circulation process for the following 4 1/2 weeks — for the production of surplus-value, it comes to exactly the same thing.

If instead we look at the turnover periods, Capital I has turned over 5 2/3 × £450 = £2,550, and Capital II has turned over 5 1/6 × £450 = £2,325 — so the total capital has turned over 5 5/12 × £900 = £4,875. That is because the turnover figure for the total capital is the sum of what I and II each actually turned over, divided by the sum of what I and II were each advanced.

It is worth noting that Capital I and Capital II, even if they stood independent of each other, would still only be different independent parts of the social capital advanced in the same

sphere of production. So if the social capital in this sphere consisted of nothing but I and II, then the very same calculation that applies here to the two parts, I and II, of one private capital would hold for the turnover of the social capital in that sphere. Carried further, every part of the total social capital invested in a particular sphere of production can be worked out the same way. And in the end, the turnover figure for the entire social capital equals the sum of the capital turned over across the various spheres of production, divided by the sum of the capital advanced across those spheres of production.

It is also worth noting: just as, within this one private business, Capital I and Capital II strictly speaking have different turnover years — since Capital II's turnover cycle starts 4 1/2 weeks later than Capital I's, so Capital I's year runs out 4 1/2 weeks earlier than Capital II's — so too the different private capitals within the same sphere of production start their businesses at quite different points in time, and so complete their year's turnover at different times of year as well. The same averaging calculation we used above for I and II is enough here too, to reduce the turnover years of the various independent parts of the social capital to one unified turnover year.

This time, the working periods and turnover periods of capital I and capital II cross each other instead of neatly following one after another. And here, unlike the case we just worked through, some capital gets released and freed up along the way.

Even so, two things still hold exactly as before. First, the number of working periods that the whole advanced capital goes through equals the value of the year's output from both capital parts, divided by the whole advanced capital. Second, the turnover number of the whole capital equals the sum of the two turned-over amounts, divided by the sum of the two advanced capitals. Here too, we have to treat the two parts of capital as though each turned over completely independently of the other.

Let's assume again that £100 has to be advanced every week to keep production running. The working period lasts 6 weeks, so it needs £600 of advance each time — this is capital I. The circulation period lasts 3 weeks, so the whole turnover period, as before, is 9 weeks. A capital II of £300 steps in during that three-week circulation period of capital I. If we treat the two as independent of one another, the table for the year's turnover looks like this:

The production process runs without a break all year, at the same scale throughout. In this picture, capital I and capital II stay completely separate from each other. But to show them as separate like this, we had to tear apart the real crossings and interminglings between them — and that changes the turnover number too. According to the table above, capital I turns over 5 2/3 × £600 = £3,400, and capital II turns over 5 × £300 = £1,500, so the whole capital turns over 5 4/9 × £900 = £4,900.

But this isn't right, because, as we'll see, the real production and circulation periods don't fully match those of the table above — a table whose whole point was to make capital I and capital II look independent of each other.

In reality, capital II doesn't have its own working period and circulation period separate from capital I's. The working period is 6 weeks, the circulation period 3 weeks. Since capital II only amounts to £300, it can only cover part of a working period — and that is exactly what happens. At the end of week 6, a product worth £600 goes into circulation, and flows back as money at the end of week 9. So at the start of week 7, capital II goes into action and covers the needs of the next working period for weeks 7 to 9. But by our assumption, at the end of week 9 that working period is only half finished. So at the start of week 10, the £600 of capital I that has just flowed back goes into action again, and its £300 covers what's needed for weeks 10 to 12. That completes the second working period. Now there is a product worth £600 in circulation, due to flow back at the end of week 15; but alongside it, £300 — the amount of the original capital II — has been released, and can go to work in the first half of the next working period, weeks 13 to 15. Once that period is over, the £600 flows back again: £300 of it is enough to reach the end of the working period, and £300 stays released for the one after.

It runs like this:

First turnover period: weeks 1 to 9.

First working period: weeks 1 to 6. Capital I, £600, is at work.

First circulation period: weeks 7 to 9. At the end of week 9, £600 flows back.

Second turnover period: weeks 7 to 15.

Second working period: weeks 7 to 12.

First half: weeks 7 to 9. Capital II, £300, is at work. At the end of week 9, £600 flows back as money — this is capital I.

Second half: weeks 10 to 12. £300 of capital I is at work. The other £300 of capital I stays released.

Second circulation period: weeks 13 to 15.

At the end of week 15, £600 — half from capital I, half from capital II — flows back as money.

Third turnover period: weeks 13 to 21.

Third working period: weeks 13 to 18.

First half: weeks 13 to 15. The released £300 goes into action. At the end of week 15, £600 flows back as money.

Second half: weeks 16 to 18. Of the £600 that flowed back, £300 is at work; the other £300 stays released again.

Third circulation period: weeks 19 to 21, at the end of which £600 again flows back as money. In this £600, capital I and capital II are now fused together — indistinguishably, with no way to tell which part came from which.

This gives us eight full turnover periods of a £600 capital (I: weeks 1–9; II: 7–15; III: 13–21; IV: 19–27; V: 25–33; VI: 31–39; VII: 37–45; VIII: 43–51), running to the end of week 51. But weeks 49 to 51 fall within the eighth circulation period, so during those weeks the released £300 has to step in and keep production going. That gives us the turnover for the year as follows: the £600 has completed its circuit eight times over, making £4,800. Add to that the product of the last 3 weeks (49–51), which has only covered a third of its 9-week circuit — so it counts toward the turnover total for only a third of its value, that is, £100. So if the year's product over 51 weeks comes to £5,100, the capital that has actually turned over is only £4,800 + £100 = £4,900. The whole advanced capital of £900 has therefore turned over 5 4/9 times — a shade more than in the first case.

In this example, we assumed a case where the working time is two-thirds and the circulation time one-third of the turnover period — so the working time is a simple multiple of the circulation time. The question now is whether the release of capital we've just found still happens when that is not the case.

Let's take a working period of 5 weeks, a circulation time of 4 weeks, and a capital advance of £100 per week.

First turnover period: weeks 1 to 9.

First working period: weeks 1 to 5. Capital I, £500, is at work.

First circulation period: weeks 6 to 9. At the end of week 9, £500 flows back.

Second turnover period: weeks 6 to 14.

Second working period: weeks 6 to 10.

First stretch: weeks 6 to 9. Capital II, £400, is at work. At the end of week 9, capital I — £500 — flows back as money.

Second stretch: week 10. Of the £500 that flowed back, £100 is at work. The remaining £400 stays released for the next working period.

Second circulation period: weeks 11 to 14. At the end of week 14, £500 flows back as money.

Up through the end of week 14 (weeks 11 to 14), the £400 released above is at work. £100 out of the £500 that then flows back completes what's needed for the third working period (weeks 11 to 15), so that once again £400 is released for the fourth working period. The same thing repeats in every working period: at its start it finds £400 on hand, enough to cover the first 4 weeks. At the end of the 4th week, £500 flows back as money, of which only £100 is needed for the last week, while the remaining £400 stays released for the next working period.

Let's also take a working period of 7 weeks, with a capital I of £700, and a circulation time of 2 weeks, with a capital II of £200.

Then the first turnover period runs from week 1 to week 9: the first working period takes weeks 1 to 7, with an advance of £700, and the first circulation period takes weeks 8 to 9. At the end of week 9, the £700 flows back as money.

The second turnover period, weeks 8 to 16, encloses the second working period, weeks 8 to 14. Of that, the need for weeks 8 and 9 is covered by capital II. At the end of week 9, the £700 above flows back; of that, £500 gets used up by the end of the working period (weeks 10 to 14). That leaves £200 released for the next working period. The second circulation period runs weeks 15 to 16; at the end of week 16, £700 flows back again. From here on, the same thing repeats in every working period. The capital needed for the first two weeks is covered by the £200 released at the close of the previous working period; at the end of the second week of that period, £700 flows back — but the working period only has 5 weeks left to run, so it can only use up £500. So £200 always stays released for the next working period.

So it turns out that in our case — where the working period is longer than the circulation period — a sum of money capital always gets released at the close of every working period, no matter what, and it is exactly the same size as capital II, the capital advanced for the circulation period. In our three examples, capital II was £300 in the first, £400 in the second, £200 in the third; and correspondingly, the capital released at the close of the working period was £300, £400, and £200 each time.

Let's go back to assuming a turnover period of 9 weeks: 3 weeks working period, needing £300 of available capital (call it Capital I). The circulation period is 6 weeks long. To cover those 6 weeks we need an extra £600 — but we can split that into two more capitals of £300 each, one for each working period. So now we have three capitals of £300 apiece. At any moment, £300 is at work in production while £600 is out in circulation.

This is the exact mirror image of Case I — the only difference is that now three capitals take turns instead of two. The capitals never cross or tangle with one another; you can follow each one separately right through to the end of the year. And just as in Case I, no capital is released at the end of a working period. Capital I is fully laid out by the end of week 3, flows all the way back by the end of week 9, and starts working again at the beginning of week 10. Capital II and Capital III behave the same way. Because the handover between them is regular and complete, no release of capital happens at all.

Here is how the total turnover works out. Capital I: £300 × 5⅔ = £1,700. Capital II: £300 × 5⅓ = £1,600. Capital III: £300 × 5 = £1,500. Total capital: £900 × 5⅓ = £4,800.

Now take a case where the circulation period is not an exact multiple of the working period — say, a working period of 4 weeks and a circulation period of 5 weeks. The matching capitals would then be Capital I = £400, Capital II = £400, and Capital III = £100. We'll only work through the first three turnovers.

Here the capitals do tangle a little, because Capital III has no working period of its own — it only lasts one week — and that week falls inside the first working week of Capital I. But in return, at the end of the working period of both Capital I and Capital II, an amount equal to Capital III, £100, gets released. Here's why. Capital III fills the first week of the second (and every later) working period of Capital I, and at the end of that week the whole of Capital I, £400, flows back. That leaves only 3 weeks of the working period still to cover, needing only £300. So the £100 that gets freed up is exactly enough for the first week of the working period that follows straight after for Capital II. At the end of that week, the whole of Capital II, £400, flows back too — but the working period still running can only absorb £300, so once again £100 is left released. And so it continues. So capital gets released at the end of a working period whenever the circulation time is not a simple multiple of the working period — and the amount released is exactly equal to whatever is needed to cover the excess of the circulation period over a working period, or over a multiple of working periods.

In every case we've looked at, we assumed that both the working period and the circulation time stay the same all year round in whatever business we're considering. That assumption was necessary if we wanted to work out how circulation time affects turnover and how much capital has to be advanced. That in reality this often doesn't hold so strictly — or doesn't hold at all — changes nothing about the point being made.

Throughout this whole section we've only looked at the turnovers of circulating capital, not of fixed capital — simply because the question at hand has nothing to do with fixed capital. Instruments of labour and the like only count as fixed capital to the extent that they last longer in use than one turnover period of the fluid capital — that is, to the extent that the time they go on serving in repeated labour processes is longer than a turnover period of fluid capital, equal to some whole number, n, of such turnover periods. Whether the total time made up by these n turnover periods is long or short, the part of productive capital that was advanced as fixed capital for that whole stretch is not advanced again during it. It just goes on functioning in its old physical form. The only difference is this: depending on how long each working period within each turnover period of fluid capital happens to be, the fixed capital hands over a bigger or smaller share of its original value to that working period's product; and depending on how long the circulation time of each turnover period is, that value-share flows back in money form faster or slower. What we're dealing with in this section — the turnover of the circulating part of productive capital — follows from the very nature of that part of capital. Fluid capital used up in one working period cannot be used again in a new working period until it has completed its own turnover: turned into commodity-capital, then into money-capital, then back into productive capital. So to run a second working period straight on from the first, fresh capital has to be advanced and turned into the fluid elements of productive capital — enough of it to fill the gap created by the circulation time of the fluid capital advanced for the first working period. That is exactly why the length of the working period of fluid capital affects the scale on which the labour process can run, and how the advanced capital has to be divided up, or topped up with fresh portions. And that is precisely what this section set out to examine.

So, from what we've worked through so far, the following results hold:

A. The different portions that capital must be split into — so that one part is always in its working period while the others are out in their circulation period — take turns with each other, like separate independent private capitals, only in two situations. First, when the working period equals the circulation period, splitting the turnover period into two equal halves. Second, when the circulation period is longer than the working period but is still an exact whole-number multiple of it — so the circulation period equals n working periods, with n a whole number. In both these situations, no part of the capital successively advanced is ever released.

B. But in every other case — first, wherever the circulation period is longer than the working period without being an exact multiple of it, and second, wherever the working period is longer than the circulation period — a part of the total fluid capital is constantly and periodically released at the end of each working period, from the second turnover onward. And the amount released is: the part of the total capital advanced to cover the circulation period, when the working period is longer than the circulation period; and the part of capital needed to cover the excess of the circulation period over a working period (or over a multiple of working periods), when the circulation period is longer than the working period.

C. It follows that for the total social capital — considered with respect to its circulating part — the release of capital must be the rule, and the mere handover of successively functioning portions of capital the exception. Because the working period being exactly equal to the circulation period, or the circulation period being exactly a whole-number multiple of the working period — this kind of neat proportionality between the two parts of the turnover period — has nothing whatsoever to do with the nature of the matter, and so, taken as a whole, can only happen as an exception.

So a very sizeable part of the social circulating capital that turns over several times a year will, over the course of the year's turnover cycle, periodically sit there in the form of released capital.

It's also clear that, with everything else staying the same, the amount of this released capital grows as the scale of the labour process grows — as the scale of production grows — in other words, as capitalist production develops generally. In the case under B.2, this is because the total advanced capital grows; in the case under B.1, it's because as capitalist production develops, the circulation period gets longer, and so does the turnover period, in those cases where the working period has no regular ratio to the two periods.

Take the first case: say we have to lay out £100 a week. For a six-week working period that's £600; for a three-week circulation period that's £300; together £900. Here, £300 is constantly released. But if instead we lay out £300 a week, then the working period needs £1,800 and the circulation period £900 — so what gets periodically released is £900, not £300.

D. A total capital of, say, £900 has to be split into two parts — as before, £600 for the working period and £300 for the circulation period. The part actually laid out in the labour process is thereby cut by a third, from £900 down to £600, and the scale of production is cut by a third along with it. The remaining £300, on the other hand, only serves to keep the working period running without a break, so that £100 can be laid out in the labour process in every single week of the year.

Looked at in the abstract, it seems to make no difference whether £600 works for 6 × 8 = 48 weeks (giving a product of £4,800), or whether the whole £900 is laid out in the labour process for 6 weeks and then sits idle for the 3-week circulation period; in that second case it would work 5⅓ × 6 = 32 weeks out of the 48 (product = 5⅓ × 900 = £4,800), and lie idle for 16 weeks. But that is only how it looks in the abstract. Set aside the extra wear on the fixed capital during those 16 idle weeks, and the extra cost of labour, which has to be paid for the whole year even though it only works part of it — a regular interruption of the production process like this is simply incompatible with running modern large-scale industry at all. Continuity of this kind is itself a productive force of labour.

Let's look more closely at the capital that gets released — or, more precisely, suspended, since that is the truer word for what is happening. A significant part of it must always take the form of money-capital. Stay with the example: a working period of 6 weeks, a circulation period of 3 weeks, an outlay of £100 a week. Halfway through the second working period, at the end of week 9, £600 flows back, of which only £300 is needed to cover spending for the rest of the working period. So at the end of the second working period, £300 of that sum is released. What state is this £300 in? Suppose one-third has to be laid out on wages and two-thirds on raw and auxiliary materials. Of the £600 that has flowed back, £200 is therefore in money form for wages, and £400 is in the form of productive stock — elements of the constant, circulating productive capital. But since only half of that productive stock is actually needed for the second half of working period II, the other half sits for 3 weeks as stock surplus to a whole working period. The capitalist knows that of this £400, he needs only half — £200 — for the current working period. So it will depend on market conditions whether he turns this £200 straight back, wholly or partly, into surplus productive stock, or holds it, wholly or partly, as money-capital while he waits for better conditions. On the other hand, it goes without saying that the part to be laid out on wages — £200 — is kept in money form. The capitalist cannot warehouse labour-power the way he can raw material once he has bought it: he has to put it to work in the production process and pay for it at the end of the week. So of the £300 released, at any rate £100 of it will, in any case, take the form of released money-capital — money-capital not needed for the working period. The capital released in the form of money-capital must therefore be at least equal to the variable capital laid out on wages; at most it can amount to the whole of the released capital. In reality it constantly fluctuates between this minimum and this maximum.

This money-capital, released purely by the mechanism of the turnover movement — alongside the money-capital released by the gradual return of fixed capital, and the money-capital needed in every labour process for variable capital — is bound to play a significant part once the credit system develops, and at the same time is bound to form one of that system's foundations.

Suppose, in our example, that the circulation time shrinks from 3 weeks to 2. Say this is not the normal state of things but the result of, say, brisk trade or shorter payment terms. The £600 of capital laid out during the working period now flows back a week earlier than it needs to — so for that week it is released. And, as before, £300 (part of that £600) is released at the midpoint of the working period, but now for 4 weeks instead of 3. So the money market has £600 released for one week, and £300 released for 4 weeks instead of 3. Since this does not happen to just one capitalist but to many, at different times, in different lines of business, more disposable money-capital appears on the market as a result. If this state of affairs lasts, then, wherever it is possible, production will be expanded; capitalists working with borrowed capital will make less demand on the money market, which eases that market just as much as an increased supply would; or, finally, the sums which…

Sums that turn out to be excess to this whole mechanism are, in the end, simply thrown straight onto the money market as available money-capital.

As a result of the contraction of the circulation time from 3 to 2 weeks, and therefore of the turnover period from 9 to 8 weeks, one ninth of the capital originally laid out simply isn't needed any more: the six-week working period can now be kept running just as steadily with £800 as it previously was with £900. A portion of the value of the commodity-capital — £100 — once turned back into money, therefore stays in that state as money-capital, without functioning any further as part of the capital advanced for the production process. While production continues on the same scale and under otherwise unchanged conditions, such as prices, the value-sum of the capital advanced falls from £900 to £800; the remaining £100 of the value originally advanced is set free in the form of money-capital. As such, it enters the money market and forms an additional part of the capitals functioning there.

This shows how a plethora of money-capital can arise — and not only in the sense that the supply of money-capital is greater than the demand for it. That is always only a relative plethora, the kind that occurs, for instance, in the "melancholy period" that opens the new cycle after a crisis ends. There is also a plethora in a different sense: a definite part of the advanced capital-value becomes superfluous for running the whole social reproduction process — circulation included — and is therefore set free in the form of money-capital. This is a plethora that arises while the scale of production and prices both stay exactly the same, purely through the contraction of the turnover period. The mass of money in circulation — larger or smaller, whatever its size — has had not the slightest influence on this.

Now suppose the reverse: the circulation period lengthens, say from 3 weeks to 5. Then already at the next turnover, the return of the advanced capital happens 2 weeks too late. The last part of this working period's production process cannot be carried on by the mechanism of the advanced capital's own turnover. If this state of affairs lasted longer, then just as expansion of the production process could occur in the previous case, contraction of it — of the scale on which it is carried on — could occur here. But to keep the process going at the same scale, the advanced capital would have to be increased, for the whole duration of this lengthening of the circulation period, by two-ninths, that is £200. This additional capital can only be drawn from the money market. If the lengthening of the circulation period holds for one or several major branches of business, it can therefore put pressure on the money market, unless that effect is cancelled out by a counter-effect from elsewhere. Here too it is plain and obvious that this pressure, like that plethora before it, had not the slightest thing to do with any change either in the prices of commodities or in the mass of the circulating medium on hand.

Finishing this chapter for the press was no small trouble, Engels writes. However sure Marx's footing was as an algebraist, he never felt at home with plain numerical calculation, and especially not with the commercial kind — even though a thick bundle of notebooks exists in which he worked through every sort of commercial calculation himself, in many worked examples. But knowing the individual methods of calculation is by no means the same thing as being practised in a merchant's everyday, working reckoning, and so Marx tangled himself up in these turnover calculations to the point that, alongside what was simply unfinished, a good deal that was wrong or self-contradictory came out of it. In the tables printed above, Engels says he has kept only what was simplest and arithmetically correct — mainly for the following reason.

The uncertain results of this laborious reckoning led Marx, in Engels's view, to give an actually rather unimportant point far more weight than it deserved. He means what Marx calls the "release" of money-capital. The real state of affairs, on the assumptions made above, is this:

Here is Engels's own account of it: whatever the ratio in size between the working period and the circulation time — that is, between Capital I and Capital II — once the first turnover is complete, the capital needed for one working period, a sum equal to Capital I, flows back to the capitalist in money form at regular intervals, each as long as the working period.

Engels's first example: if the working period is 5 weeks, the circulation time 4 weeks, and Capital I is £500, then a sum of £500 flows back each time — at the end of week 9, 14, 19, 24, 29, and so on.

Engels's second example: if the working period is 6 weeks, the circulation time 3 weeks, and Capital I is £600, then £600 flows back each time — at the end of week 9, 15, 21, 27, 33, and so on.

And Engels's third example: if the working period is 4 weeks, the circulation time 5 weeks, and Capital I is £400, then £400 flows back each time — at the end of week 9, 13, 17, 21, 25, and so on.

Engels adds: whether, and how much, of this returned money is surplus to the current working period — that is, released — makes no difference. It is assumed that production goes on uninterrupted on its current scale, and for that to happen the money has to be there, has to flow back, whether "released" or not. If production is interrupted, the release stops too.

In other words, Engels grants that release of money does happen — that latent, merely potential capital in money form does form — but it happens under all circumstances, not only under the special conditions spelled out in the text, and it happens on a larger scale than the text assumes. With respect to the circulating Capital I, the industrial capitalist finds himself, at the end of every turnover, in exactly the position he was in when he first set up the business: he has the whole sum in hand again, all at once, even though he can only turn it back into productive capital gradually.

What matters in the text, Engels says, is the demonstration that a considerable part of industrial capital must always exist in money form, and that an even larger part must temporarily take on money form as well. These additional remarks of his, at most, reinforce that demonstration. — F. E.

Up to now we've held prices steady and let the turnover period stretch or shrink instead. Now let's do the opposite: hold the turnover period and the scale of production steady, and let prices change — the price of raw materials, of auxiliary materials, and of labour, or of the first two of these. Say the price of raw and auxiliary materials, and wages, falls by half. In our example that means £50 a week instead of £100, and £450 of advanced capital for the nine-week turnover period instead of £900. £450 of the capital value that used to be advanced is now set free — for the moment, as money capital — while production keeps running at the same scale, with the same turnover period split up the same way as before. The year's output is still the same quantity of goods, but its value has fallen by half. None of this was produced by a speeding-up of circulation, or by a change in the amount of money circulating — though this price change does go along with a shift in the supply and demand for money capital.

But the causation actually runs the other way. The fall in the value, or price, of the elements of productive capital by half would, first of all, mean that business X — still running at the same scale as before — only needs half as much capital value advanced to it, and so only throws half as much money onto the market, since business X advances that capital value first in the form of money, as money capital. The amount of money thrown into circulation would have shrunk, because the prices of the elements of production fell. That was the first effect.

But there is a second effect. Half of the capital value originally advanced — £900, so £450 — used to pass through the forms of money capital, productive capital, and commodity capital in turn, and at any one moment was split across all three forms at once. That £450 is now set free from business X's circuit and enters the money market as an additional block of money capital — it acts on the market as something extra. This freed £450 acts as money capital not because it has become surplus money no longer needed to run business X, but because it is part of the original capital value — so it goes on functioning as capital, rather than being spent as a mere means of circulation. The next way for it to act as capital is to be thrown onto the money market as money capital. Or, alternatively, the scale of production could be doubled instead (fixed capital aside). The same £900 of advanced capital could then run a production process twice the size.

If, on the other hand, the prices of the fluid elements of productive capital rose by half, £150 a week would be needed instead of £100 — so £1,350 instead of £900. £450 of additional capital would be needed to keep the business running at the same scale, and depending on the state of the money market, this would put a greater or lesser strain on it. If all the capital available on the market were already spoken for, competition for available capital would intensify. If some of it were lying idle, that idle capital would now be drawn into use.

There is also a third possibility. With the scale of production fixed, the speed of turnover unchanged, and the price of the fluid elements of productive capital unchanged, the price of what business X actually sells — its own products — could still rise or fall.

Say the price of the goods business X delivers falls. Then the price of its commodity capital of £600 — the amount it constantly throws into circulation — drops, say, to £500 (the surplus value locked up in that commodity capital doesn't come into it here). A sixth of the value of the advanced capital fails to flow back from the circulation process; it is simply lost there. But since the value, or price, of the elements of production stays the same, this reflux of £500 is only enough to replace five-sixths of the £600 capital constantly employed in production. So £100 of additional money capital would have to be laid out to keep production going at the same scale.

The reverse case: if the price of business X's products rose, the price of its £600 commodity capital might rise to, say, £700. A seventh of that price — £100 — doesn't come out of the production process at all; it was never advanced there, but comes purely from the circulation process. Yet only £600 is needed to replace the elements of production. So £100 is set free.

Why the turnover period shortens or lengthens in the first case, why the prices of raw material and labour rise or fall in the second, or why the prices of the delivered products rise or fall in the third — none of that belongs to the inquiry we've been conducting.

But this does belong here:

Under the assumptions of our example, shortening the circulation period means a ninth less total capital needs to be advanced — bringing it down from £900 to £800, and setting free £100 of money capital.

Business X still delivers the same six-weekly output worth £600 as before, and since work carries on without a break through the year, it delivers the same total quantity over 51 weeks, worth £5,100. So nothing changes in the quantity or the price of the product it throws into circulation, nor in the timing of when it puts that product on the market. Yet £100 has been set free, because shortening the circulation period now lets the process run on only £800 of advanced capital instead of the previous £900. The £100

of set-free capital exists in the form of money capital. But it is by no means the part of the advanced capital that has to keep functioning permanently in the form of money capital. Suppose that of the £600 advanced fluid capital I, four-fifths — £480 — is constantly laid out on materials of production, and one-fifth — £120 — on wages: £80 a week on materials, £20 a week on wages. Capital II, £300, must be split the same way: four-fifths, £240, for materials, one-fifth, £60, for wages. Capital laid out on wages must always be advanced in money form. As soon as the commodity product worth £600 is reconverted into money — sold — £480 of it can be turned into materials of production (into productive stock), but £120 keeps its money form, held back to pay six weeks' wages. That £120 is the minimum of the returning £600 capital that must always be renewed and replaced in the form of money capital — it must always be there, permanently, as the part of the advanced capital functioning in money form.

Now, of the £300 that is periodically set free every three weeks — itself splittable the same way into £240 of productive stock and £60 of wages — shortening the circulation time throws £100 clean out of the turnover mechanism, set free in the form of money capital. Where does the money for this £100 of money capital come from? Only a fifth of it is money capital that was already being periodically set free within the ordinary turnovers. But four-fifths of it — £80 — has already been replaced by additional productive stock of the same value. So how does this additional stock get converted into money, and where does the money for that conversion come from?

Once the circulation time has actually shortened, only £400 of the £600 is reconverted into productive stock, instead of £480. The remaining £80 is held fast in its money form, and together with the £20 for wages makes up the £100 of set-free capital. It's true that this £100 comes out of circulation by way of the purchase of the £600 commodity capital, and is now withdrawn from circulation because it isn't laid out again on wages and materials of production — but it's worth remembering that, in money form, it is back in exactly the form in which it was originally thrown into circulation. Originally, £900 in money was laid out on productive stock and wages. To carry out the very same production process, only £800 is needed now. The £100 set free in money form by this now forms a new money capital looking for somewhere to invest — a new component of the money market. It's true that this £100 was already found, periodically, earlier on, in the form of freed money capital and of additional productive capital — but those earlier, temporary states were themselves a condition for carrying out the production process at all, for keeping it continuous. Now they are no longer needed for that, and so they form new money capital and a new component of the money market — even though they are neither an additional element added to the existing social stock of money (since they already existed when the business began, and it was the business itself that threw them into circulation) nor a newly accumulated hoard.

The £100 is now genuinely withdrawn from circulation, in the sense that it's the part of the advanced money capital no longer employed in this same business. But that withdrawal is only possible because the conversion of commodity capital into money, and of that money back into productive capital — the movement W′–G–W — has been sped up by a week, and so the circulation of the money active in that process has been sped up too. It is withdrawn because it is no longer needed for the turnover of capital X.

We've been assuming the advanced capital belongs to the person using it. If it were borrowed instead, nothing about this would change. With the circulation time shortened, he would now need only £800 of borrowed capital instead of £900. The £100 handed back to the lender still forms £100 of new money capital exactly as before — only now it sits in Y's hands instead of X's, rather than disappearing. Or again: suppose capitalist X gets his £480 worth of materials of production on credit, so that he only has to advance £120 in money himself, for wages. He would now need £80 less of materials on credit — so that £80 becomes excess commodity capital sitting with the capitalist who extends the credit — while capitalist X himself would have set free £20 in money.

The extra production stock has now shrunk by a third. It used to be 240 pounds — four-fifths of the 300 pounds of extra capital II — and now it is only 160 pounds. That means enough stock for two weeks instead of three. It gets renewed every two weeks instead of every three, but each renewal is smaller, covering only two weeks' worth instead of three. So purchases — on the cotton market, say — happen more often and in smaller lots. The same total amount of cotton is drawn out of the market, because the total output hasn't changed. What changes is how that withdrawal is spread over time: in smaller pieces, over more separate purchases. Take an example: suppose the period is three months in one case and two months in the other, with a yearly cotton consumption of 1,200 bales. In the first case, the amounts sold are:

By contrast, in the second case:

So the money laid out on cotton comes back in full a month later — in November instead of October. Suppose the shorter circulation time, and so the shorter turnover, releases a ninth of the capital advanced — 100 pounds — in the form of money capital. Suppose further that this 100 pounds is made up of 20 pounds that periodically sat idle as surplus money for the weekly wage, and 80 pounds that periodically existed as one week's worth of surplus production stock. Then, for that 80 pounds, the smaller surplus stock on the manufacturer's side is matched by a larger stock of goods sitting with the cotton dealer. The very same cotton stays that much longer on the dealer's shelves as unsold goods as it stays shorter on the manufacturer's shelves as raw material waiting to be used.

Up to now we assumed that the shorter circulation time in business X came from X selling the goods faster, or being paid faster — or, where credit is involved, from a shorter payment term. That kind of shortening traces back to a shorter sale of the goods: the turn of commodity-capital into money-capital, the first phase of the circuit. But the shortening could just as well come from the second phase, the turn of money into goods, and so from a simultaneous change — in the working period, or in the circulation time — of the businesses that supply capitalist X with the raw materials for his circulating capital.

Take an example: suppose cotton or coal used to spend three weeks travelling from where it is produced or stored to the site of capitalist X's works. Then X's minimum production stock has to last at least three weeks until fresh supplies arrive. While the cotton and coal are in transit, they cannot yet serve as means of production. Instead, at that stage, they form

the raw material of the transport industry — goods being worked on by the capital employed there — and, from the point of view of the coal producer or the cotton seller, commodity-capital still on its way to market. Now suppose transport improves and the journey shrinks to two weeks. The production stock can then shrink too, from a three-week stock to a two-week one. That frees up the 80 pounds of extra capital that had been tied up to cover it, and the 20 pounds tied up for wages as well, because the 600 pounds of turned-over capital now flows back a week sooner.

Or, going the other way: suppose the working period of the business that supplies the raw material gets shorter (earlier chapters gave examples of this), so that the raw material can be replaced more often. Then the productive stock can shrink, and the gap between one replenishment and the next can shorten too.

If instead the circulation time — and with it the turnover period — gets longer, then extra capital has to be advanced to cover it. It can come out of the capitalist's own pocket, if he has spare capital lying around. But that spare capital will usually be tied up in some other form, as part of the money market, and freeing it up — selling shares, pulling out deposits, and so on — has its own indirect effect on the money market. Or he has to borrow it. The part of the extra capital needed for wages must, under normal conditions, always be advanced as money capital, and it is here that capitalist X puts his own share of direct pressure on the money market. For the part to be spent on raw materials, direct pressure is unavoidable only if he has to pay cash. If he can get the materials on credit, there is no direct effect on the money market at all, because the extra capital is then advanced straight into the production stock rather than first appearing as money capital. If his creditor turns around and throws the bill he got from X back onto the money market — has it discounted, say — then the effect reaches the money market indirectly, at second hand. But if the creditor instead uses that bill to settle some debt of his own falling due later, then this extra capital never touches the money market at all, neither directly nor indirectly.

We had assumed a total capital of 900 pounds, laid out as 720 pounds in raw materials and a fifth, 180 pounds, in wages.

Suppose the price of raw materials falls by half. Then the six-week working period only needs 240 pounds instead of 480, and the second, supplementary capital only needs 120 pounds instead of 240. Capital I shrinks from 600 pounds to 240 plus 120, that is 360 pounds. Capital II shrinks from 300 pounds to 120 plus 60, that is 180 pounds. The total capital of 900 pounds shrinks to 360 plus 180, that is 540 pounds. So 360 pounds is set free.

This set-free capital, now idle and looking for somewhere to go on the money market, is nothing but a slice of the very same 900 pounds originally advanced as money capital. It has become surplus to requirements — assuming the business stays at the same scale rather than expanding — because the price of the raw materials it is periodically turned back into has fallen. Now, if that price fall were not owed to some accidental circumstance (an especially rich harvest, an oversupply, and so on) but to a rise in the productive power of the branch that supplies the raw material, then this money capital would be an outright addition to the money market — to the capital available in the form of money capital generally — because it would no longer form an integral part of the capital already at work in the business.

Now consider a price fall: part of the capital is simply lost and has to be replaced by a fresh advance of money capital. This loss to the seller may be made good again by the buyer. Directly, if the product's market price fell only because of some passing turn of the market and afterward climbs back to its normal level. Indirectly, if the price change was really caused by a change in value — a change that reaches back onto the old product — and if that product then enters another branch of production as an element of production and frees up capital there. In either case, the capital that X lost, and for whose replacement he now leans on the money market, may be supplied to him by his business partners as fresh extra capital. When that happens, nothing new is created — one capital has simply been transferred to another.

Now consider a price rise: a portion of capital that was never advanced at all gets pulled out of circulation as a gain. It was never part of the sum he actually laid out for production, so — as long as production is not expanded — it counts as set-free money capital. Since the prices of the product's elements were fixed before it went to market as commodity-capital, a genuine change in value could be behind this rise, working backward — say the raw materials went up in price afterward. If so, capitalist X gains twice over: on the product still circulating as commodity-capital, and on the production stock he already holds. That gain would hand him exactly the extra capital he now needs, at the new, higher prices of the raw materials, to keep his business running.

Or the price rise may only be temporary. Whatever becomes necessary as extra capital on capitalist X's side then turns up as capital set free on the other side, wherever his product serves as an element of production for some other branch of business. What the one has lost, the other has gained.