Macro Notes 1: Aggregate Demand

1.1  Goods Market
We are now moving into macroeconomic theory. The theory we will start with is called the Income-expenditure model. This model looks at the Goods Market (or the Market for Goods and Services). This is just the first piece of the picture of how the macroeconomy works -- we will keep adding to this model as the semester goes on.

1.2  Aggregate Income and Aggregate Output
Aggregate Output is the total amount of output produced and supplied in the economy in a given period. Aggregate Income is the total amount of income received by all factors of production in an economy in a given period. The two of them are always equal at any period of time, so we can refer to both of them as aggregate income, and use the symbol Y to describe them (can you explain why the two are always equal?).

1.3  Identities, Behavioral Equations, and Equilibrium Conditions
We need to distinguish between an identity and an equation before we can proceed with our analysis. An identity is a statement that is true by definition at all times. Thus, for example, when we say that Yd = C + S that is an identity, since it is always true - there is nothing else people can do with their disposable income.

An equation is a description of a specific type of relationship, and does not have to be true at all times. In economics, we distinguish between two types of equations:

Behavioral equations or functions. These tell us what people would like to do, and how they would like to behave (whether they actually do manage to achieve their desired behavior met depends on the economy, and so we cannot assume that behavioral equations are true at all times).

Equilibrium equations tell us what relationship must exist if everybody is to manage to satisfy their desires (as described in the behavioral equations) at the same time. But while an equilibrium equation or condition can tell me what has to happen if everybody is to be able to meet their desired behavior simultaneously, I do not have any guarantee that the economy is actually at that position! Thus the equilibrium equation is only true for those situations when everybody actually does manage to satisfy their desired behavior. In such a situation, there is no tendency for things to change (since everybody manages to meet their desired behavior, and so no one finds that they cannot meet their decisions and tries to change things)--which is why it is called an equilibrium.

An Equilibration process tells me how the economy actually moves to a situation where everybody manages to meet their desired behavior (given from the behavioral functions). That is, it tells me how the economy actually reaches equilibrium. If the equilibration process works, then every time an economy is out-of-equilibrium, things will change, until the economy reaches equilibrium.

(Here is a simple example from micro: "quantity supplied = quantity demanded" is an equilibrium condition. The equations for the demand and supply functions (curves on a graph) are behavioral equations. Suppose that price is lower than equilibrium. In this case quantity demanded will exceed quantity supplied, and not all consumers will get as much of the good as they want. In this case all consumers will not "achieve their desired behavior," as we said above, and the equilibrium condition is not satisfied. But we assume that the market will not remain long in this situation, because firms will raise prices in response to apparent excess demand for these goods. Only in equilibrium will both buyers and sellers satisfy their behavioral equations.)

1.4  Aggregate Expenditure
The income-expenditure model zeroes in on a problem that firms face in a modern capitalist economy: how much to produce? In other words, how much demand can forms expect for the goods they make? If there are enough expenditures, then firms are covering all the incomes they have to pay out. So, to find out if there are enough expenditures, we have to look at the desires of different groups of people to purchase goods and services. This will give us the behavioral equations for each of these groups.

Note the categories of expenditure we had identified earlier: C, I, G, X and M.

To keep the model simple, for now we will omit the Rest of the World.

(Remember that what we started with a national income identity, where we said that GDP is always identically equal to C+I+G+X-M. But that was based simply on the actual amount of expenditures on C, I G, X and M found in the economy. It was not based on the desired spending on C, I, G, X and M. Thus, what we had before was an identity, which may or may not have been a level of GDP where everybody managed to meet their desired levels of expenditure. The key to this difference is the fact that "I" contains not just planned acquisition of capital goods by firms, but also unanticipated changes in their inventories of goods.)

1.5 Aggregate Consumption Behavior
How much do consumers wish to spend? We will focus on the relationship between aggregate income Y (remember this is also the same thing as aggregate output) and consumption C.

(C here is not the same thing as your demand from the demand and supply analysis in micro. That was the demand for a single good, which depended on its price relative to the price of other goods, taste or preferences for one good over another, and so on. What we have here is the total level of consumption expenditure on all goods by all households in the economy.)

Now note that the actual consumption households undertake depends on their disposable income, because they don't have any choice about paying taxes. So consumption and savings will be functions of disposable income, or (Y-T).

Since whatever is not consumed must be saved, as soon as we specify a consumption function we have necessarily specified a savings function.

To keep things simple, we are going to specify consumption as a linear (straight line) function:

C = a + bY

in which "a" represents some basic level of consumption people will undertake regardless of income (assume they dip into savings if their income is zero) and "b" represents the amount of each additional dollar earned people will spend on goods and services. (In the language of analytic geometry, "a" is the "intercept" and "b" is the "slope" of the line.)

This "b" has a special name: the Marginal Propensity to Consume (MPC). In economic terms, it tells the additional amount of aggregate consumption that the members of the economy will desire to undertake, for each additional dollar of income they receive.

The MPC is always positive (since when people earn more, they will consume more).

The MPC is also less than 1. That is we assume that some part of each extra dollar earned is saved. That gets us to the next point, We know from our savings identity that in all circumstances

S = Y - C

So, once we know our consumption function, we can always derive the relationship between Y and S. We can also easily figure out the Marginal Propensity to Save. Since every extra dollar earned is either saved or consumed,

MPC + MPS = 1

E.g. if my MPC is .75, I spend seventy-five cents of each extra dollar earned on goods and services, so I must be saving the remaining quarter. Hence my MPS is .25.

(Let's introduce some shorthand notation here. We'll use "" to mean "change in." In that case we can say that MPC = C/Y and that MPS = S/Y )

1.6  Investment Behavior
We examined how consumers decide on their level of expenditures. Let us now examine how firms decide on their level of expenditures.

Now remember that in our GDP identity, we had a category called "I" for investment. But that was simply the total amount of actual investment that the firms ended up undertaking, regardless of whether they desired to have this level of investment or not. That is, the actual I we used in our GDP calculations included everything that ended up with firms including their unsold goods ("inventory") regardless of whether this was a desired level of investment.

Here, we are looking at what firm owners want to spend, so we are looking at the behavioral equation for investment. This we will call Ip (or planned investment). Ip essentially refers to purchases of physical or productive capital, such as planned purchases of tractors, buildings, plant machinery, and so on. (If a firm wants to build up its inventories we should also include that inventory change in planned investment, but to keep things simple we can ignore that possibility.)

In addition, however, the actual investment "I" includes unplanned inventory buildup (or decline): additions to inventory because firms were not able to sell the amount they thought they would be able to. This means that if there is any unplanned investment, firms are not meeting their planned or desired investment behavior.

OK, so how do we specify the planned investment function? Very simply. For now, we will assume that Ip does not vary with Y. In other words we take Ip as given.

1.7  Government Purchases
This is even easier. We will assume that government chooses its desired level of purchases, so we will also take G as given.

Since G is under the control of policymakers, we can also use this model to explore the consequences of a change in the amount of government purchases. (Ip, by contrast, is under the control of individual capitalists and we assume the government has no power to tell them what to do.)

(This is a good place to introduce a couple of terms:

exogenous: determined outside the model

endogenous: determined inside the model

Here G is exogenous. On the other hand C is endogenous, because it's determined inside the model, by the consumption function.)

1.8  Aggregate Expenditure and Equilibrium
We now have C, Ip, and G. Since we are assuming a closed economy, we forget about X and M. That means we have all the information we need about the planned level of total (aggregate) expenditure in the economy:

Planned Aggregate Expenditure = C + Ip + G

Equilibrium occurs when the amount of output that firms wish to sell (which is the same as the amount of income in the economy) Y, is the same amount as households and firms and government wish to buy. When that happens, everybody's desired decisions are met, and there is no tendency for change in the economy.

Thus our equilibrium condition is: Y = C + Ip + G

Here is a good point to be sure we have this business about planned and unplanned investment (and about identities and equilibrium conditions) under control. Suppose that firms make too much stuff. That means that:

Y > C + Ip + G

Because they still have to pay incomes to the workers who make the stuff. But we already stated as an identity that:

Y = C + I + G

Is this a contradiction? No! Remember that our broad category "I" is the sum of planned investment (Ip) plus inventory changes. So if firms make $10 billion worth of goods but C + Ip + G = $9.9 billion, then firms will end up with $100 million of extra unsold goods, in other words their inventories will rise an unanticipated $100 million.

When we add that inventory increase to Ip to get the total I, then the identity stated above holds. But this is not equilibrium, because firms' total investment exceeds their planned or intended investment: I > Ip. So the identity holds even when we are not in equilibrium.

You can work out the corresponding situation when I < Ip.

Another way of looking at the same equilibrium condition is to ask: when will the amount of desired expenditures by everybody absorb exactly all of Y? C, the largest part of Y, is uncomplicated. But T and S do not automatically convert themselves into spending. To put it formally, we know from our (closed-economy) identities both that

Y = C + I + G

and that

Y = C + S + T

which means that

C + I + G = C + S + T


I + G = S + T

Since in equilibrium I = Ip, we can now re-express the equilibrium condition in our macroeconomy as:

Ip + G = S + T

In other words when the part of individual/household income that is not spent by individuals/households exactly equals the planned spending of firms and the spending of government, we are in equilibrium, with no further tendency to change.

If you are given a consumption function and the pre-set amounts of G and Ip, you can solve for the equilibrium level of Y by writing down the equilibrium condition Y = C + Ip + G and then substituting in the consumption function for C, and the pre-set amounts of Ip and G. This will give you an expression you can solve for Y. Since it's easy to make a calculating mistake in this process, get used to checking your answer by substituting the equilibrium Y you have just found into the consumption function to get a value for C, and then adding it to the values for Ip and G, to see if you get C+Ip+G=Y.

1.9  Equilibration Process
How does our economy actually reach this point? We know that the economy is not always in equilibrium. Some people would argue that it never achieves complete equilibrium. How does the economy move from a situation of disequilibrium toward its equilibrium?

This is a critical question. You cannot assume that some sort of macro god descends from the sky and tells firms how much to make. You can not assume that the economy spontaneously "finds" its equilibrium position. Equilibrium here means a position toward which the macroeconomy tends to move. (Similarly in a micro model the equilibrium price was the one toward which the market would tend to move - if it was higher it would tend to fall, if lower it would tend to rise - all because of plausible actions undertaken by firms.)So if you cannot explain the tendency, if you cannot explain why an out-of-equilibrium economy tends to move toward equilibrium, then you don't understand the model.

Suppose C + Ip + G < Y. Then output/income is greater than desired expenditures. Firms find that they have unintended increases in unsold inventories. What will the firms do when they cannot sell all their output? Will they continue to produce as much as they did before? No. They cut back on output and hence income falls.

That's the core idea. Let's follow the whole story.

When Y > C + Ip, Y decreases because of the responses of firms.

When income falls, what happens to C? Does it stay as high? No. When income falls, consumers find that they have less income and so they spend less. (Note that while consumers spend less, they do not decrease their consumption by the full amount of the drop in income because MPC is less than 1.)

So when C falls, total planned expenditures (C + Ip + G) fall too. But because MPC<1, C+Ip+G does not fall quite as much as Y falls. So what? Well, the fact that Y fell more than C+Ip+G means that the gap between them has narrowed. So we are at least part way along in the story about how our initial problem (Y > C + Ip + G) is resolved.

If it's still true that Y > C + Ip + G, then firms will cut output again. If it happens that firms guessed right and Y = C + Ip + G, then nothing further will happen: we are at equilibrium, at rest.

If firms cut output too much, or if our story starts with too little output, then

we have a situation in which Y < C + Ip. In this case inventories will fall, not rise, so that inventory change will be negative and I will fall short of Ip. Firms, seeing this, will expand output and hence Y will rise. And so on.

Can you see that the MPC being less than 1 is very important for the ability of the economy to reach equilibrium?

1.10  The Multiplier
Now we know how the economy moves toward equilibrium, and we can find out what the equilibrium level of income in an economy will be. But what happens to equilibrium income when one of the exogenous factors in expenditures change? In particular, what happens if we change government purchases or taxes?

(Remember that you should never assume that equilibrium is rapidly or easily achieved. If the economy is in equilibrium and we then change something like G, it is not going to immediately jump to the new equilibrium, but will go through a process like the one described in the previous section. So what we are really asking here is: "If we change an exogenous factor like G, what is the new center of gravity toward which the economy will tend?")

Starting with an original equilibrium income level, we find that if one of the exogenous components (like Ip) increases, this will increase total expenditures by that amount. But immediately, this sets of our equilibrating process.

Next, firms will recognize the additional demand for goods and raise output to meet that extra demand. As a result, Y will rise.

But this is not the end of the story!

The fact that Y begins rising means that incomes are going up. As Y rises, C must rise too. Each extra dollar of Y raises C by that dollar times the MPC (remember that? If not go back to section 5 above).

As C rises, that represents new demand for goods, and as firms meet that demand Y rises even more. Then C rises, Y rises, C rises, Y rises etc. This ripple effect is why equilibrium Y rises more than just the initial increase in Ip or G. Or why it falls more, if Ip or G fall.

How much more? If the MPC is 0.9 then a $1 rise in G means:

$1.00 in extra G leads to $1 in extra Y which leads through the MPC to

$0.90 in extra C which leads to .90 in extra Y which leads to

$0.81 in extra C which leads to .81 in extra Y which leads to

$0.729 in extra C which leads to .729 in extra Y which leads to

$0.656 in extra C which leads to .656 in extra Y which leads to ...


(down to very very small numbers)

If you add up all of this series, it so happens that you will get a total rise in Y of $10.

If the MPC is 0.5, then a $1 rise in G means:

$1.00 in extra G leads to $1 in extra Y which leads through the MPC to

$0.50 in extra C which leads to .50 in extra Y which leads to

$0.25 in extra C which leads to .25 in extra Y which leads to

$0.125 in extra C which leads to .125 in extra Y which leads to

$0.0625 in extra C which leads to .0625 in extra Y which leads to ...


(down to very very small numbers)

If you add up all of this series, it so happens that you will get a total rise in Y of $2. If you have dealt with this sort of infinite series in math class, you'll recognize what's going on mathematically. If not, don't worry. The key thing you need to recognize is that the larger the MPC, the bigger each successive ripple in the pond is: with the MPC = 0.5 each the ripples dies away pretty fast, while with MPC = 0.9 they're a lot bigger.

In fact the multiplier = 1/(1-MPC) in this model. But in this course, don't trouble yourself with memorizing the formula. Know the basic idea.

The same process happens in reverse if G or Ip falls. Suppose you were starting at equilibrium. Then something happened to planned investment - say that firm owners became despondent about their future prospects for sales increases, and cut Ip.

If G and T remain unchanged, then Y and C will fall until a new equilibrium is reached.

(Here's another way to think about what will happen, and to think about the math. Since nothing is happening with G or T, then if we started with

Ip + G = S + T

then once we achieve the new lower equilibrium, S will have fallen exactly as much as Ip was cut. Right? So the change in S (at the new equilibrium) will equal the change in Ip that started this disturbance. Now follow carefully:

1. The multiplier answers the question: what is the total change in Y if there is a given change in Ip (or G)?

2. We just said that the change in S will be the same amount as the change in Ip (once the new equilibrium is reached).

3. And we already know that the MPS = S/Y (Remember "" means "change in")

4. So if S = Ip, then MPS = Ip/Y too, right?

5. And if MPS = Ip/Y, then 1/MPS = Y/Ip (we invert each side)

6. Y/Ip means "the change in Y per dollar change in Ip" which is what the multiplier is. So the multiplier = 1/MPS.

7. And since MPS = 1-MPC, the multiplier also = 1/(1-MPC)

1.11  Fiscal Policy
By changing G, we have already been doing fiscal policy. Let's deal with the subject more carefully.

Government Purchases are all the direct expenditures on final goods and services by the Government. We will refer to this as G.

Taxes are all the income and sales and other taxes the government takes out of the income flow. Transfer payments are all the transfers of income like social security, unemployment compensation, and so on that the government gives to households. Note that this is not direct expenditure on goods and services by the government but is a flow to households. Net Taxes is the net amount of taxes less transfer payments that the government takes out of the circular flow. We will refer to this as T. (To keep it simple we'll usually just talk about lowering or raising taxes, but you can see that raising transfer payments would change Yd just as much as lowering taxes)
So, we have Y = a + b (Y-T) + I + G

By changing G or net taxes T the government can change equilibrium income (Y). This is called fiscal policy. Although states, cities, and even counties tax and spend in the United States, for purposes of this course we will focus on the federal government. Following the Constitution, the President proposes a budget but it is the U.S. Congress that decides on taxing and spending.

We have already shown how to use our simple model to evaluate the effects of changing G: equilibrium Y rises or falls by the amount of the change in G times the multiplier. What about T? Note that taxes and transfers do not affect expenditures directly. They affect expenditures by affecting the amount of disposable income, and so they work their effects through C. So suppose government raises taxes by $100 million. That lowers disposable income by $100 million, which lowers consumption by $100 million multiplied by the marginal propensity to consume. So if the MPC is .9, then the first effect on aggregate demand that the $100 million tax increase has is a $90 million drop in C. After that, the rest of the multiplier story works the same as before - Y down $90 million, C down another $81 million, Y down $81 million etcetera etcetera.

But the first step in the (net) tax multiplier story was just a little different: if instead of raising taxes $100 million we had lowered government purchases $100 million, then that $100 reduction on G, because it is a direct component of aggregate demand, would have brought about a reduction in Y of $100 million, followed by C going down $90 million and so on.

Lowering G $100 million:

$100 million in less G leads to $100 million in less Y which leads through the MPC to

$90 million in less C which leads to $90 million in less Y which leads to

$81 million in less C which leads to $81 million in less Y which leads to


(down to very very small numbers)

All these changes will sum to a drop in Y of $1 billion.

Raising T $100 million: The higher T means a drop in C of $90 million.

$90 million in less C leads to $90 million in less Y which leads to

$81 million in less C which leads to $81 million in less Y which leads to


(down to very very small numbers)

All these changes will sum to a drop in Y of $900 million.

So the difference between raising taxes $100 million and lowering government purchases $100 million is that the first impact on aggregate demand is different. Changing G means directly changing part of AD, while a change in T has to work through the MPC before it has its first direct effect on AD.

So while G produces Y in the full amount of the multiplier, T produces (negative)Y in the amount of the multiplier times the MPC.

In formula terms, since the multiplier for G is 1/(1-MPC), the multiplier for T will be -MPC/(1-MPC. It gets a minus sign because if T goes up Y goes down.

Finally, note that the model we have is very simple -- we are assuming that the Government assesses a fixed amount of taxes, and changes that fixed amount. A more realistic model would assess a tax rate as some proportion of Y.

Now we come to a textbook chestnut: the "balanced budget multiplier." Suppose we raise (net) taxes and raise government purchases by the same amount. Or we lower taxes and lower government purchases by the same amount. What is the net effect on the economy?

You already have a sense of the answer, from our comparison of the effects of similar changes in G and T above. Because a change in G affects AD fully, while a change in T affects AD only in slightly diminished form (by changing C first through the MPC), changing spending is just a little more powerful than changing taxes.

And in fact, you already know enough to tell exactly how much change in Y will be provoked by a matched change in G and T. Let's raise both G and T by $100 million, and keep the MPC = .9 from the previous example.

We already know that by raising T $100 million we get a drop in C of $90 million.

$90 million in less C leads to $90 million in less Y which leads to

$81 million in less C leads to $81 million in less Y which leads to


(down to very very small numbers)

All these changes will sum to a drop in Y of $900 million.

But that's not the whole story, because we also raised G $100 million. And by doing that:

$100 million in more C leads to $100 million in more Y which leads to

$90 million in more C which leads to $90 million in more Y which leads to

$81 million in more C which leads to $81 million in more Y which leads to


(down to very very small numbers)

All these changes will sum to a rise in Y of $1 billion.

So the total effect of raising T by $100 million was that Y fell $900 million. The total effect of raising G $100 million was a rise in Y of $1 billion. The net combination of these two effects is that Y rose, but only by $100 million.

And in fact, in this simple model the balanced budget multiplier is always exactly 1.

(If algebra makes you happy, you can get this result by adding up the two abstract formulas:

1/(1-MPC) as the multiplier for G, and -MPC/(1-MPC) as the multiplier for T. Add them and you get (1-MPC)/(1-MPC), which is 1.)

1.12  Counter-cyclical and Pro-cyclical Policies
When the economy is in a recession, the government can increase G and/or decrease T to increase demand and income. When the economy is booming and inflationary pressures start to grow in the economy, the Government can decrease G and increase T. If the budget is normally more or less in balance, then this means that the government runs deficits in recessions, and surpluses in booms. This should stabilize the level of aggregate expenditure and income in an economy. When the government does this, it is called counter-cyclical policy. Essentially the government is trying to damp down swings in Y.

If these swings in Y are part of a normal "business cycle" in which periods of intense capital investment alternate with periods in which firms buy relatively few new capital goods, then it's especially easy to see the rationale for counter-cyclical G: If firms' intended investment (Ip) falls, that's a component of AD and Y will tend to fall. In that case, in theory, G can be increased to make up for the fall in Ip. In real life, this is hard because it may take a while to actually figure out that Ip is dropping, and the political process of approving changes in G or T may drag on for long enough that by the time fiscal policy is actually changed, Ip has risen again. In this case your intended counter-cyclical policy might actually end up being a pro-cyclical policy, amplifying rather than damping the changes in Ip.

Counter-cyclical policy would also lower G when Ip rises, to reduce booms. You might wonder why anyone would want to do this - aren't booms good? The most often-heard arguments are (a) that a boom sets up conditions for a painful crash by encouraging over-investment (too much Ip, so that it collapses once firms realize they have bought too many machines) and (b) that overly-rapid growth provokes rapid inflation.

1.13  Automatic Stabilizers
Note that in our simple economy, we have assumed that G and T are fixed, and don't depend on income Y. But in a more sophisticated model, transfer payments and taxes in particular will change as Y changes.

If tax revenues are a percentage of income, then as Y rises taxes will rise by themselves.

If transfers like unemployment compensation rise when people lose their jobs and fall when employment rises, then when Y rises transfers fall, and when Y falls transfers rise.

So since net taxes (T) represent total taxes minus transfer payments, it follows that T will rise when Y rises and fall when Y falls. Note that this amounts to a counter-cyclical policy as described in the previous section, but that it's automatic - it requires no extra decision by government to do this. This kind of countercyclical policy is also pretty rapid.

1.14  Deficits and Debt
What happens when the government runs a deficit - that is when G>T? It borrows money. Another way of saying the same thing is that it sells securities (IOUs). If G>T, the size of the difference (G-T) - which is how much has to be borrowed - is called the deficit.

But suppose the government already owes money from previous deficits. Then this year's deficit adds to the total debt of the government. So the federal debt is the total amount owed by the federal government, while the deficit os the amount this debt rises in a single year.

In other words the debt is the cumulative total of all past deficits.

So while recent deficits have been around $200 billion, the total federal debt is approaching $5 trillion.

1.15  Further Notes
You have heard a lot of discussion in recent years about the federal deficit and debt.

Some of this debate has been interesting, and reasonable people can take very different positions on taxing, spending, and deficits. But unfortunately a lot of the discussion has been based on the fallacy that national debt is just like personal debt. Personal debt has to be paid off by a certain point: I might take out loans to go to college, but I won't be able to continue borrowing forever (lenders know I have a finite earning life), and at some point I have to pay it all back. But the U.S. government has an infinite life. Additionally, because it has the power to tax nobody will worry about its ability to pay back in the future. So government can keep "rolling over" its borrowing: issuing new securities as the old ones come due. Of course it still has to pay interest, but the "principal" - the amount of the original borrowing - never has to be repaid.

This does not mean that we have discovered some kind of magic beans. Government borrowing does have consequences and they can be, arguably, bad. But to think about those consequences you have to think in real terms: what is the change in real, physical, output and the allocation of that output that will result from running a fiscal deficit?

Let's tick off some (not all) of the reasons that deficits might harm or help.

Deficits might be useful for:

1. Countercyclical policy: as argued above, raising G or lowering T (either by deliberate policy or through automatic stabilizers) can help reduce the severity of a recession. In real terms, this would mean that there is less lost output during recessions - when output drops that means that workers and machines that could be making stuff are idle.

2. Capital expenditures: Businesses borrow all the time to buy capital equipment. Suppose government wants to build a highway system. Some economists argue that if the highway system will raise future incomes and hence tax revenues over the future, it makes sense to borrow the money to build the highways, and then tax incomes to repay the borrowing. Of course, this means increasing taxes after the highway system is built, and people won't like that. So there's a built-in temptation to keep on borrowing. (When people argue that it's "their money" and that the government has no right to it, they ignore the fact that their ability to make an income depends partly on government spending on their education, on the roads they use, on the military that defends their interests, on the police and judiciary that keeps them safe, and so on.) In real terms, all this amounts to saying is that setting up a "capital budget" would make it easier to identify whether G was going into things that raised everyone's Y in the future.

Note that these are two arguments for borrowing for specific things, but not for running a large or rapidly-growing debt. What are the reasons for objecting to deficits? There are two main ones:

1. The budgetary burden of higher interest payments: As the total debt rises, the annual interest payments go up too. (You can see that in your data.) If those payments rise faster than taxes (which will rise as overall Y rises), then interest payments make up a large part of federal outlays every year. The result of this is that taxpayers pay interest to people who hold the government's debt. In the aggregate, the effect is a wash: some people have less income from taxes, others have more from interest payments. But if government debt is held mainly by rich people, while the tax burden is more evenly distributed, then having a large debt may tend to transfer command over resources from poorer people to wealthier ones - a real effect.

(A related argument has to do with what happen if foreigners own a lot of the debt. The government can't tax foreigners. That means it will pay the foreigners interest in dollars, and the foreigners can use those extra dollars to buy our stuff (without giving us any of their stuff in exchange). So we might end up having to run a trade surplus if foreigners stop buying new U.S. debt.)
2. Crowding out: If G>T, government borrows. In order to attract savings, government may have to bid against businesses that are trying to borrow money for capital investment projects (remember how Ip is financed in our simple model). When government bids against capitalists for savings, it may have to offer a higher interest rate, and at the higher interest rate capitalists may then borrow less and undertake less Ip. So on this argument, if G rises without a rise in T, then government "crowds out" private sector borrowing, and goods/services that would have gone to private firms now flow to government - a real effect.

©1998 S. Charusheela and Colin Danby.