Thinking in a Macro Way

To understand aggregate behavior, it is necessary to start at the aggregate or macro level of analysis rather than the individual or micro level. This is because there are certain relationships that must hold, by definition, at the aggregate level. These relationships are specified in macroeconomic accounting identities.

Reasoning at the individual and micro level is also necessary to understand many aspects of the economy, including the aggregate economy, but the micro reasoning must be consistent with the relationships holding at the macro level. In considering the aggregate economy, macro reasoning takes precedence over micro reasoning. If a micro argument is inconsistent with known macro relationships, it is the micro argument that must be rejected.

A major strength of modern monetary theory (MMT) is that it starts from relationships that must hold true at the aggregate level, and then ensures that any behavioral assumptions that go beyond what immediately follows from these aggregate relationships (and so are contestable) are fully consistent with the accounting identities. The same can be said of Keynes, Kalecki, Marx, and also to some extent the classical political economists (for instance, in their analysis of broad class behavior). MMT is the result of applying macro consistent reasoning in a way that takes account of institutional features of modern monetary systems.

But while macro consistency has guided MMT economists, and in varying degrees Keynes, Kalecki, Marx, and the classical political economists, it has not guided neoclassical macroeconomists, who have tried to start at the individual level and then simply add up (aggregate) individual results to obtain a picture of the aggregate economy. Starting from the micro and reasoning up to the macro, as the neoclassicals do, results in many fallacies of composition. Neoclassical general equilibrium theorists have themselves uncovered various aggregation problems. They run so deep that basically neoclassical macroeconomics cannot say anything of much interest about the aggregate economy without either assuming a ‘single commodity world’ or a ‘single consumer world’. Otherwise, there is no sound basis for aggregating the individual decisions of investors, producers, consumers, etc.

None of this is to suggest that there is only one valid theory of the aggregate economy. There is more than one way to conceive of economic behavior that is consistent with the macroeconomic accounting identities. However, if a theory is not macro consistent, it cannot be a valid theory of the aggregate economy.

Having said this, starting at the aggregate level is actually very revealing in terms of the types of behavioral assumptions we might want to make. In the remainder of this post, I want to provide an example of reasoning in a macro way. The approach begins with incontestable macroeconomic relationships and then employs contestable behavioral assumptions in an effort to go beyond what follows immediately from the macro identities. As a consequence, the arguments presented here are to a large degree contestable, but they are consistent with the macro relationships. In other words, I am giving one example of macro consistent reasoning.

One of my favorite examples of the insights possible with macro reasoning is due to Kalecki. For simplicity, we can begin by considering a pure closed private-sector economy. Government and external sectors will be included shortly.

Kalecki starts with two accounting identities that must hold for the period under consideration:

(1) Y = C + I

(2) Y = W + P

Here, Y is real output or income, C is private consumption, I is gross private investment, W is aggregate wages, and P is aggregate profits realized in the period.

In Marxian fashion, Kalecki then divides consumption into capitalist consumption out of profits, CP, and worker consumption out of wages, CW, and assumes that workers, in aggregate, consume all they earn: CW = W (this assumption can be relaxed without altering the basic insights). The first identity becomes:

(1′) Y = W + CP + I

Setting (1′) equal to (2) and rearranging gives:

P = CP + I

I find this expression very interesting. It may seem strange at first glance, but the expression tells us that capitalists’ own spending decisions as a class (CP + I) must equal their total profit (P). The reason for this is that workers’ wages return to the capitalists as a class when workers spend on consumption items, whereas capitalists’ expenditures go to other capitalists, and so to the capitalist class as a whole. Kalecki expresses this as “capitalists get what they spend and workers spend what they get”.

Now consider causation. Do profits determine the amount of capitalist expenditure, or do capitalist expenditures determine aggregate profit? Kalecki observes that capitalists can choose what they spend in the current period but not what they earn, implying that causation runs from capitalist expenditures to aggregate profit.

Rearranging the expression gives:

P – CP = I

Since workers have been assumed to spend all they earn in aggregate, the left-hand side of this expression represents private saving, S. It is that part of aggregate profits not consumed by capitalists. So:

S = I

Since Kalecki has already observed that capitalists choose what they spend, not what they earn, this implies that causation runs from investment to saving. That is, capitalist investment determines private saving.

Keynes, of course, in a similar analysis showed that at the aggregate level, income is determined by investment, and that saving is a function of income. Thinking at the aggregate level enabled us to deduce that, in contrast to pre-Keynesian thinking, investment is not determined by saving (e.g. through interest-rate adjustments), but rather investment determines saving through income adjustments. If anyone still doubts, bear in mind that the orthodox explanation of investment and saving was shown to be invalid – a fallacy of composition – in the capital debates. If we start at the macro level, we would never have posited this theory in the first place.

Kalecki’s analysis can be extended to gain insight into some of the factors influencing the average (economy-wide) realized rate of profit, r, which roughly speaking can be represented as aggregate profit divided by the amount of fixed capital, K, tied up in production. That is, r = P/K. (For Marx, the denominator should also include outlays on raw materials and wages, but if the turnover period for these items is short relative to the accounting period under consideration – e.g. a year – the impact on the economy-wide rate of profit is small, and doesn’t alter the basic issues under discussion.)

Drawing on Kalecki’s analysis, we have:

r = (CP + I) / K

In this expression, gross private investment, I, will add not only to profit in the numerator, but also to the amount of fixed capital, K, in the denominator to the extent that gross investment exceeds depreciation of fixed capital. Here, ΔK = I – δ, where δ is depreciation. So investment adds to aggregate profit and aggregate fixed capital.

To see the significance of this, consider an extreme example in which CP = 0, I = 100 each year and there is zero depreciation. If K = 1000 initially, the rate of profit in successive years will be 100/1000, 100/1100, 100/1200, … . That is, there will be a tendency for the rate of profit to decline as capital intensity increases. (The tendency will be mitigated to the extent that depreciation occurs.) This is similar to Marx’s tendency for the rate of profit to fall, except expressed in price rather than value terms and applied to realized profit rather than produced surplus value. In the case where all surplus value produced is realized, the two measures — produced surplus value and realized profit — coincide. For Marx, this tendency for the rate of profit to decline will continue until a crisis causes the value of K to collapse. Once prices of the components of fixed capital fall, the value of K in the denominator will be reduced, and the rate of profit revived.

It is interesting to note that, technically, there is an “out” which could help capitalists avoid a fall in the realized rate of profit. The “out” is capitalist consumption, Cp. Provided there is excess capacity, which is the normal condition in a capitalist economy, capitalists could increase their consumption relative to their investment. To the extent that they did this, profit (the numerator) would increase without causing the capital stock (the denominator) to expand. The decline in the rate of profit would be avoided.

From Marx’s (contestable) perspective, there is a limit to this method of sidestepping the tendency for the rate of profit to fall, because competition compels capitalists to reinvest some of the previous period’s profit in the current period in an effort to improve productivity (compete on cost) and increase productive capacity (maintain market share). Marx’s argument here is an example of a behavioral assumption that is contestable (i.e. not the only possible behavioral assumption) but fully consistent with the relationships that we know must hold at the aggregate level. Kalecki’s perspective is similar: current investment determines current aggregate profit; current aggregate profit, to the extent it influences future profit expectations, will have some influence on next period’s investment.

Once the government sector is included, Kalecki’s analysis reveals something that I think is highly significant. Depending on the monetary system, fiscal policy could be used to moderate or even completely offset any tendency for the realized rate of profit to fall. In a commodity-backed monetary system or common currency arrangement, this would be difficult. But in modern monetary systems — those involving flexible exchange-rate fiat currencies — governments have such a capacity if they choose to use it.

To see this, we can introduce the government and external sectors into Kalecki’s analysis. The basic identities become:

(1”) Y = C + I + G + NX

(2′) Y = W + P + T

where Y is income, G is government expenditure, T is tax revenue, and NX is net exports.

Making the same assumptions and following the same procedure as we did in the case of a pure private-sector economy gives:

P = CP + I + BD + NX

where BD = budget deficit = G – T.

Finally, the average rate of profit becomes:

r = (CP + I + BD + NX) / K

With the extra sectors included, there are now other possible “outs” to avoid a falling rate of profit. The most important “out” is the government’s capacity to run budget deficits. In fact, for the world economy as a whole, NX = 0, and the only “out” at the global level would appear, from a Marxist perspective, to be the government’s capacity to run budget deficits.

If there is commodity-backed money or a fixed or pegged exchange rate, and the government is compelled to balance the budget overall (not each year, but over the business cycle), the budget cannot influence the average rate of profit, because BD = 0 on average. Even so, the government could smooth the decline of the rate of profit by using countercyclical fiscal policy. However, for open economies, there is a problem for trade-deficit nations, because fiscal policy is constrained by exchange-rate imperatives. This limits the government’s capacity to neutralize the effects of the falling rate of profit.

From an MMT perspective, a flexible exchange-rate fiat-money system has an important advantage in this respect. Budget deficits could be used to increase the rate of profit without creating any debt burden in the future. MMT shows that the government is at liberty to net spend with its own fiat currency to the extent that real resources are idle. Doing so can boost the rate of profit by ensuring productive capacity is more consistently utilized (i.e. ensuring the numerator, P, rises in step with the denominator, K). Actually, the government can also boost productive capacity in ways that do not increase K, or can even reduce the value of K. Public investment in education as well as science and technology enables improvements in productivity, and hence reductions in prices of the elements of fixed capital. Public investment in infrastructure expands productive capacity without adding to the value invested in K by capitalists. The demand entailed in this deficit expenditure, whether public consumption or public investment, adds to the numerator of the expression for the average rate of profit without adding to the denominator.

By giving governments the fiscal capacity to overcome any tendency for the realized rate of profit to fall, fiat money opens up a variety of options for society. Some options will be preferred by those on the right, others by those on the left.

If it is felt that the business cycle and capitalist boom and bust (connected to Marx’s tendency for the rate of profit to fall) is beneficial as a spur to innovation and restructuring, the private sector could be left more or less to its own devices. Right-libertarians or conservatives might prefer this approach. Depending on their preferences concerning the poor and unemployed, this approach could be combined with a job-guarantee scheme that pays minimum wage or a basic-income guarantee. This would allow the private sector to operate more or less without government involvement (or influence on the rate of profit), while cushioning workers from the worst effects of the business cycle. If the cost of the job-guarantee scheme or basic-income guarantee were matched by taxation (a balanced budget), there would be no impact on the rate of profit, and crises could be left to play their functional role of reviving the rate of profit. (In an open economy, the budget would need to be in deficit to the extent that there is a trade deficit, but small government could still be achieved by opting for a budget deficit that is low tax, low spend.) I do think there is a big challenge at the moment for those who want small government. The private-debt problem needs to be addressed before the private sector can sustainably become the engine for economic activity. One way to address this difficulty would be to arrange for an orderly write down of private debt.

For liberals, moderates, social democrats, etc., that want a mixed economy with a bigger role for government, fiat money under a flexible exchange-rate regime makes it technically possible for the government to use fiscal policy to offset the tendency for the rate of profit to fall. The government would be boosting the rate of profit by adding to demand and capitalists’ revenue without increasing their capital costs. The smoothing of movements in the rate of profit would reduce the intensity of crises. Essentially, spurts of private investment that expanded the capital stock relative to non-investment demand would lower the rate of profit in ensuing periods unless other components of demand (which do not increase K) were stepped up to absorb the capacity increases brought about by the private investment. The government could use its fiscal policy for this purpose.

Moving towards socialism, another option would be to start doing more things in the public sector rather than using government expenditure to prop up the rate of profit in the private sector. For instance, government expenditure could be used to provide ‘free’ goods and services to the general community (free in terms of access, not resource expenditure). Increased taxation would be necessary to create space for the expanded public-sector activity.

Irrespective of society’s ideological make-up, the basic point is that fiat money opens up paths to a better society, however that is conceived. With fiat money, capital doesn’t have to be determining of what we do (to the extent we retain it, we are master), whereas commodity-backed money systems and fixed or pegged exchange-rate systems make capital the determiner of everything (we are its slaves). I think it is a pity Marx died before getting to his in-depth studies of government and the external sector, because he may well have discerned the role commodity-backed money plays in narrowing society’s options and, in contrast, the opportunities opened up by fiat money.

There is no one correct theory of where we should be heading as a society. But our contestable theories should at least be consistent with macroeconomic accounting identities.

9 thoughts on “Thinking in a Macro Way

  1. A third Peter here – and agree, this is a very good post. For novices like myself it really puts things in perspective. Will be exploring more of your posts, thanks a lot!

  2. Peter C,

    I agree with Peter D: this post is dynamite. The whole site, not only the history of economic thought posts, is terrific.

    I’ll recommend it to my friends (who are also MMTers, but without much knowledge about economics in general).

  3. In your later analysis you appear to have made up your own definition of GDP (it should be “Y = C + I + G + NX”). Taxes are not subtracted from this.

    I also note that your falling rate of profit (and subsequent conclusions) only work in an ideal world of a fixed rate of investment and capital which never wears out. In the real world, capital (plant and equipment) needs to be replaced, so K doesn’t increase monotonically, and investment increases as capital increases (presuming the economy grows as the capital stock increases). So your conclusions appear to apply to this idealized economy, not to the real world.

  4. ThomasW: Welcome. Thanks for taking the time to comment.

    Yes, that was a confusing treatment of income, which has now been amended. In the original post, I referred to income net of tax, which meant that GDP = Y + T = C + I + G + NX and Y = C + I + G – T + NX. I should have included tax in income. I have now corrected the two identities to Y = C + I + G + NX and Y = W + P + T. The resulting expression for profit is, of course, unaltered.

    The link on Kalecki provided in the post presents the following equivalent expression for profit:

    P = Cp + I + Dg + Ee – Sw

    Here, Dg is government debt for the period (equal to the budget deficit, G – T), Ee is the net external surplus (i.e. NX), Sw is worker saving (assumed zero in Kalecki’s simpler models).

    Regarding your point about depreciation, that is addressed in the post. The numerical example sets the rate of depreciation to zero to illustrate the effect, but it is simple enough to show the rate of profit falling with depreciation occurring. Modifying the example, let gross investment be 100 each year, depreciation 50 (or whatever, provided net investment is positive). The rate of profit in successive periods will be 100/1000, 100/1050, 100/1100. Clearly, the rate of profit falls, but at a slower rate than when depreciation is ignored. This was mentioned in the post as follows:

    [T]here will be a tendency for the rate of profit to decline as capital intensity increases. (The tendency will be mitigated to the extent that depreciation occurs.)

    True, you can increase I from period to period as well, but any increase in I beyond what is already included in the numerical example above will equally add to K. Repeating the modified version of the example, but allowing I to increase by 50 each period, gives rates of profit 100/1000, 150/1100, 200/1250, …, which clearly doesn’t help the situation.

    In the absence of deficit expenditure sufficient to maintain profitability (by either maintaining a high rate of capacity utilization or using public investment to limit the growth of K relative to P), the revival of profitability requires crisis and a collapse in the prices of the components of K.

  5. Marxist tricky stuff. But this seems like a pretty important post, even if I don’t know why yet.

    #Bookmarking

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>