Krugman and Obstfeld: Government Deficit Reduction May Not Increase the Current Account Surplus

I thought it would be useful to review the twin deficits debate since Treasury has just released a paper on this topic (noted here). To do so, I am going to turn things over to two people who know this topic as well as anyone anywhere, Paul Krugman and Maurice Obstfeld. This is from the 6^th edition (2003) of their International Economics textbook. Equation (12.2) referenced below is:

S^P = I + CA – S^g = I + CA – (T – G) = I + CA + (G – T)     (12.2)

where S is saving, I investment, CA the current account, T is taxes, and G is government spending. The superscript P and g on the saving term indicate private and government saving. Here’s the Case Study from page 306-307 of the text:

Government Deficit Reduction May Not Increase the Current Account Surplus

The linkage among the current account balance, investment, and private and government saving given by equation (12-2) is very useful for thinking about the results of economic policies and events. Our predictions about such outcomes cannot possibly be correct unless the current account, investment, and saving rates adjust in a way that is consistent with (12-2). Because that equation is an identity, however, and is not based on any theory of economic behavior, we cannot forecast the results of policies without some model of the economy. Equation (12-2) is an identity because it must be included in any valid economic model; but there are any number of models consistent with (12-2).

A good example of how hard it can be to forecast policies' effects comes from thinking about the effects of government deficits on the current account. During the administration of President Ronald Reagan in the early 1980s, the United States slashed taxes and raised some government expenditures, generating both a big government deficit and a sharply increased current account deficit. Those events gave rise to the argument that the government and current account deficits were "twin deficits," both generated primarily by the Reagan policies. If you rewrite identity (12-2) in the form

CA= S^P – I – (G – T),

you can see how that outcome could have occurred. If the government deficit rises (G - T goes up) and private saving and investment don't change much, the current account surplus must fall by roughly the same amount as the increase in the fiscal deficit. In the United States between 1981 and 1985, the government deficit increased by a bit over 2 percent of GNP, while S^P- I fell by about a half a percent of GNP, so the current account fell from approximate balance to about -3 percent of GNP. (The variables in (12-2) are expressed as percentages of GNP for easy comparison.) Thus, the twin deficits prediction is not too far off the mark.

The twin deficits theory story can lead us seriously astray, however, when changes in gov­ernment deficits lead to bigger changes in private saving and investment behavior. A good example of these effects comes from European countries' efforts to cut their government budget deficits prior to the launch of their new common currency, the euro, in January 1999. As we will discuss in Chapter 20, the European Union (EU) had agreed that no member country with a large government deficit would be allowed to adopt the new currency along with the initial wave of euro zone members. As 1999 approached, therefore, EU governments made frantic efforts to cut government spending and raise taxes.

Under the "twin deficits" theory, we would have expected the EU's current account surplus to increase sharply as a result of the fiscal change. As the table below shows, however, nothing of the sort actually happened. For the EU as a whole, government deficits fell by about 4.5 percent of output, yet the current account surplus remained about the same.

The table reveals the main reason the current account didn't change much: a sharp fall in the private saving rate, which declined by about 4 percent of output, almost as much as the increase in government saving. (Investment rose slightly at the same time.) In this case, the behavior of private savers just about neutralized governments' efforts to raise national saving!

Year	CA	SP	I	G-T
1995	0.6	25.9	19.9	-5.4
1996	1.0	24.6	19.3	-4.3
1997	1.5	23.4	19.4	-2.5
1998	1.0	22.6	20.0	-1.6
1999	0.2	21.8	20.8	-0.8

Source: Organization for Economic Cooperation and Development. OECD Economic Outlook 68 (December 2000), Annex Tables 27, 30, and 52 (with investment calculated as the residual).

It is difficult to know why this offset occurred, but there are a number of possible explanations. One is based on an economic theory known as the "Ricardian equivalence" of taxes and government deficits. (The theory is named after the same David Ricardo who discovered the theory of comparative advantage-recall Chapter 2-although he himself did not believe in Ricardian equivalence.) Ricardian equivalence argues that when the government cuts taxes and raises its deficit, consumers anticipate that they will face higher taxes later to pay off the result­ing government debt. In anticipation, they raise their own (private) saving to offset the fall in government saving. Conversely, governments that lower their deficits (thereby increasing government saving) will induce the private sector to lower its own saving. Qualitatively, this is the kind of behavior we see in Europe in the late 1990s.

Economists’ statistical studies suggest, however, that Ricardian equivalence doesn't hold exactly in practice. Most economists would attribute no more than half the decline in European private saving to Ricardian effects. What explains the rest of the decline? The values of European financial assets were generally rising in the late 1990s, a development fueled in part by optimism over the beneficial economic effects of the planned common currency. It is likely that increased household wealth was a second factor lowering the private saving rate in Europe.

Because private saving, investment, the current account, and the government deficit are jointly determined variables, we can never fully determine the cause of a current account change using identity (12-2) alone. Nonetheless, the identity provides an essential framework for thinking about the current account and can furnish useful clues.

Comments

[Winslow R.]

Krugman wrote:

"Because private saving, investment, the current account, and the government deficit are jointly determined variables, we can never fully determine the cause of a current account change using identity (12-2) alone. Nonetheless, the identity provides an essential framework for thinking about the current account and can furnish useful clues. "

Krugman uses the term,

SP – I

without defining what it is. Yes I know that SP is U.S. private sector savings and I is investment. But what is the difference between U.S. private sector savings and investment?

This is something that has been 'bothering' me for years. The Levy institute defines SP-I as the U.S. private sector savings rate of U.S. fiat dollars.

From the article below SP-I is the U.S. personal savings rate?

"The government reported last week that consumers last year spent all they earned and then some, pushing the personal savings rate into negative territory at minus 0.5 percent.

The savings rate has only been negative for a full year twice before, in 1932 and 1933, when Americans were struggling with huge job layoffs during the Great Depression. "

http://www.cbsnews.com/stories/2006/02/07/business/main1293943.shtml

It still doesn't quite seem right -0.5% compared to -2% but at least it is close.

From Mankiw:

"(7) S = I

This equation is called the saving-investment identity. Interpreting this as a model of long-run equilibrium, we are looking at equilibrium from the perspective of the market for loanable funds. The equilibrium real interest rate in the long-run must be such that the supply of loanable funds (aggregate saving) equals the demand (investment)."

http://www.georgetown.edu/faculty/dibab/econ002/read/Saving_Investment.pdf#search='S%20%3D%20I%20%20mankiw%20equilibrium'

Loanable funds = demand

Right, we can create money for any borrower out of thin air. I don't think this is what Mankiw meant, but it's right.

[Winslow R.]

The real version of the Expenditures Approach, then, says
(1) Y = C + I + G + X - M

6. Using the definition of Disposable Personal Income, we get
(3) Sp = Y - T - C

http://www.georgetown.edu/faculty/dibab/econ002/read/Saving_Investment.pdf#search='S%20%3D%20I%20%20mankiw%20equilibrium'

Here is the problem:

Sp = Y - T - C

should be in my mind

Sp = Y - T - C - I

Why doesn't spending on investment reduce/increase private sector savings? If C, consumption can be spent, why can't I, investment? Both can reduce private sector savings, Sp of financial assets if they are used to invest/consume offshore. Whether the private sector spends on C,consumption or I, investment dollars are transferred from one party to another. No financial assets are destroyed or created.

Even if Mankiw has mixed savings/spending of financial assets with savings/spending of nonfinancial assets the Sp should still include an I, investment term. Housing price increases should then be included in our national savings and Mankiw's definition. Sp should include a term for increasing asset market values such as housing.

Mark, anyone?

[Winslow R.]

Should have written:

Sp = Y - T - C + I

[Bruce Wilder]

I'm with you Winslow R. "Savings" is a conceptual mess. As far as I can tell, it has always been so. Keynes, in the General Theory, devotes the better part of two chapters to the tautological meaninglessness of S=I, as defined in the double-entry bookkeeping framework of the National accounts, but, evidently, made no lasting impression.

In a barter economy, there can be no savings, per se, only consumption and investment. In a barter economy, if you are not working, you are consuming leisure, and whatever you accumulate by deferring consumption, is an investment, in inventory, if nothing else.

"Saving" only makes sense in a money economy, where currency and financial assets exist. But, the economists are not defining "savings" as a financial act or transaction, something people do distinct from other economic activities or transactions.

If a corporate CEO presented his balance sheet, and said that his auditors could find no large correlation between changes in working capital and the ratio of owner's equity to total equities, we would recognize that he was nuts. If a financial analyst confused an increase in retained earnings with an increase in equity market valuation, we would write him off. Creating an identity from a system of double-entry bookkeeping and then using it as if it were a model of the world is every bit as confused and confusing.

[spencer]

Winslow -- private savings includes corporate or business savings that is always -- at least in the modern US economy -- much larger then private savings.

Use to be a good rule of thumb that business savings financed business investment and personal savings financed housing.

[chuck roast]

If we include I in investment terms in Sp and assume that I includes the value of housing, then the US did not have a personal savings rate of 0.5 last year. We might well have had a huge increase in personal savings. Also, this would tend to validate the Ricardian Equivlance theory.
However, unlike what we would typically call “investment”, savings in the form of housing stock, while a store of value, has no productive utility value. So, this tremendous increase in I last year, did not contribute to productivity.
In view of the post below, where it is said that one way out of the budget deficit is economic “growth,” how can this form of investment contribute to economic growth? Perhaps we can continue to cover the landscape with houses at an ever expanding rate. While we may be able to rub two houses together to make a third house we cannot export a house to balance the CA deficit, and if too many people have to re-mortgage or sell their houses to satisfy the budget deficit then this will lead to deflation at the least.

[knzn]

Winslow (et. al.): “Why doesn't spending on investment reduce/increase private sector savings?” Because when an entity invests, while it reduces its financial savings, it has an offsetting increase in physical savings. If you use your savings to purchase gold, you haven’t reduced your net savings, just changed its form. Similarly, if you use your savings to build a house or a factory, you haven’t reduced your net savings; you’ve spent money, but you’ve received something else that is still savings. To look at it from another angle, suppose you’re a grain farmer with a large silo. You can save in financial form by selling your whole crop and putting money in the bank, or you can save in physical form by keeping some of your grain in the silo. If you change your mind, sell the grain from your silo, and put the money in the bank, you haven’t increased your savings, just changed its form.

Mark, Unless you believe the Ricardian equivalence argument, the Eurpoean example given by Krugman and Obstfeld seems only to be saying that, when you cut your fiscal deficit, there may be other things happening at the same time that offset the effect on the trade deficit. It’s not at all clear that the increase in private saving was a result of the deficit cuts. The example doesn’t contradict the “twin deficits” theory at all. Similarly, if you reduce your fiscal deficit, perhaps at the same time, someone in your country will discover a wonderful new technology that draws in investment, and then the trade deficit goes up. That doesn’t contradict the “twin deficits” theory; it just says that there’s more to life than twin deficits.

[Winslow R.]

Thanks for the responses.

Bruce wrote:

"Saving" only makes sense in a money economy, where currency and financial assets exist. But, the economists are not defining "savings" as a financial act or transaction, something people do distinct from other economic activities or transactions.

I agree, though as long as financial assets are kept separate from nonfinancial savings accounting identities can be useful. Goverment deficits are the part of the identity that the government controls most directly. Foreign savings can be controlled through tariffs, private savings through various fiscal policies.

spencer wrote:

"private savings includes corporate or business savings that is always -- at least in the modern US economy -- much larger then private savings."

I don't understand. Private savings is larger than private savings?

chuck roast wrote:

"However, unlike what we would typically call “investment”, savings in the form of housing stock, while a store of value, has no productive utility value."

Economists would like us to think so? Except to raise families? This seems to be of the mindset that businesses are productive and families are unproductive justifying twisted tax policy.

"While we may be able to rub two houses together to make a third house we cannot export a house to balance the CA deficit"

I believe if a house/asset/business is purchased by a foreign entity the C/A deficit is reduced even though nothing tangible is exported except for the title.

knzn wrote:

"Because when an entity invests, while it reduces its financial savings, it has an offsetting increase in physical savings. If you use your [financial] savings to purchase [nonfinancial] gold, you haven’t reduced your net savings, just changed its form. "

Yes, but you miss the point? What if the value of those nonfinancial savings increase, say gold goes from $200/oz to $800/oz? Why does this not increase 'savings' as defined by Mankiw?

"That doesn’t contradict the “twin deficits” theory; it just says that there’s more to life than twin deficits."

If you think 'twin deficit' theory requires a direct correlation between fiscal deficits and the C/A deficit than there is a contradiction. The 'triple deficit' theory says there is a direct link between savings of financial assets. The American population/economists/politicians are confused by the the inclusion of nonfinancial assets in the calculation.

[Emmanuel]

Winslow R.,

The Mankiw textbook differs from Krugman and Obstfeld's because it calls the current account deficit S(subscript )f or "foreign" saving. Mankiw's total saving, therefore, is private + government + "foreign" saving, whereas Krugman and Obstfeld only include private and government saving in total saving then separate the current account deficit.

This is the key point: in the link that you gave, note 7 says my definitions here follow the exposition in your [Mankiw's] textbook, which is suitable for the GNP accounts. A better definition for the GDP accounts would be to define foreign-sector saving as the current account deficit. In other words, the instructor prefers the Krugman & Obstfeld definition.

I kind of prefer Krugman & Obstfeld's definition which is more conventional, but if you tease out "foreign saving" from Mankiw's definition of total saving and express it as the current account deficit, you should arrive at the same thing.

Hope this helps.

[Winslow R.]

Emmanuel wrote:

"This is the key point: in the link that you gave, note 7 says my definitions here follow the exposition in your [Mankiw's] textbook, which is suitable for the GNP accounts. A better definition for the GDP accounts would be to define foreign-sector saving as the current account deficit. In other words, the instructor prefers the Krugman & Obstfeld definition."

I don't think this solves why Mankiw doesn't include increases in values of domestic investments in U.S. private sector savings.

" The difference between GNP and GDP lies in the treatment of income from foreign sources. GNP measures the value of goods and services produced by U.S. nationals, while GDP measures the value of goods and services produced within the boundaries of the U.S. "

http://www.cals.ncsu.edu/course/are012/readings/gdp&lead.html

I would just say you have brought up the crux for the 'dark matter' controversy. I believe 'dark matter' is tied into the confusion caused by mixing financial and nonfinancial assets. Income from abroad is based on assets with 'book' not 'market' values so our foreign assets are undervalued showing huge ROI's.

[knzn]

Winslow: “What if the value of those nonfinancial savings increase, say gold goes from $200/oz to $800/oz? Why does this not increase 'savings' as defined by Mankiw?”

That is an increase in the nominal value of savings but not in the real value. An important part of the story in the real world is that changes in asset values (and goods prices) shift the savings from one owner to another, so that some people end up with more savings and others with less. However, the total real savings cannot be altered by changes in prices. If the price of gold goes up, the few owners of gold have more savings, but each of the many owners of cash has slightly less savings, in real terms (since, e.g., they can’t buy as much jewelry with their bond proceeds). Similarly, if housing prices go up, current owners have more savings, but future owners have less.

[Winslow R.]

"That is an increase in the nominal value of savings but not in the real value...However, the total real savings cannot be altered by changes in prices."

You bring up the confusion of inflation/deflation. Would you agree the 'slippage' in values between financial and nonfinancial assets is inflation/deflation?

Who is to determine when nominal savings becomes real? Will the IMF always value their gold at $40-50 /oz?

Gold falls on IMF sale concerns

"The IMF values its gold reserves at between $40 and $50 an ounce, a price that was fixed in the 1970s and is about a tenth of the metal's current market value....Bringing the book price of the gold in line with market value would boost the IMF's balance sheet, giving it more money to distribute. "

http://news.bbc.co.uk/2/hi/business/4242565.stm

[knzn]

“Would you agree the 'slippage' in values between financial and nonfinancial assets is inflation/deflation?” I wouldn’t use the term inflation/deflation, but that’s a purely semantic issue. The point is, prices change, but the total quantity of real savings does not. (Some of that savings, I should note, comes from seignorage, which is the production of a good called money.) The IMF’s gold valuation is a meaningless accounting convention; the market determines what different types of savings are worth relative to one another.

[Bruce Wilder]

gold is not unique in this respect. wealth, in general, is a matter of expectations of future prices and income, and is disconnected from the past. Wealth cannot be accumulated savings, if savings are defined exclusively in terms of past acts.

investments, even an investment in an inventory of gold, is valued in terms of market expectations of future prices and liquidity. should savings and investments be regarded as identical?

if the government runs a surplus and pays down the national debt, does it really make sense to say that private wealth is thereby "reduced"?

[Bruce Wilder]

Does any of this shed light on the thesis that the current value of the yuan-dollar exchange rate overvalues investment in China and undervalues investment in the U.S.?

[Winslow R.]

knzn wrote:

"The point is, prices change, but the total quantity of real savings does not."

I don't think you've made your point. What are 'real savings', nonfinancial assets? In the IMF's case, gold?

This is why the discussion about the 'triple deficits' always gets sidetracked. The 'triple deficit' theory is only concerned about savings of 'unit of account', fiat money not gold. Mankiw continues the confusion.

Snow is concerned about the U.S. private sector savings rate going negative - the first time since the Great Depression by the way.

The 'triple deficit' theory tells us:

1) If we push down on the fiscal deficit we may push the U.S. private sector savings even more negative.

2) If the C/A deficit increases further, which has been its current path, the U.S. private sector savings rate will go even more negative.

3) If we increase U.S. private sector savings while the gov deficit stays the same the C/A deficit will fall.

Bruce wrote:

"if the government runs a surplus and pays down the national debt, does it really make sense to say that private wealth is thereby "reduced"?"

Exactly, the 'triple deficit' theory says if the U.S. government pays down national debt, the combination of U.S. private sector and foreign sector savings of U.S. dollars would be reduced. Not 'real wealth'.

Bruce wrote:

"Does any of this shed light on the thesis that the current value of the yuan-dollar exchange rate overvalues investment in China and undervalues investment in the U.S.?"

It would overvalue foreign 'income' and undervalue foreign 'assets' in my book.

[James Kroeger]

Winslow, Wilder, et al...

You guys might be interested in some of my attacks on the presumed legitimacy of the SAVINGS = INVESTMENT identity.

In The Misunderstood Relationship Between Savings & Investment I try to use logic and empirical evidence to show why there is little more than an incidental relationship between savings levels and interest rates. The bottom line: interest rates are an EXOGENOUS variable that does not respond to changes in Savings.

In a recent blog entry I tried to provide a simplified explanation of the major flaw in the calculation of the BEA's Personal Savings Rate.

I think some of these arguments might help to put the relationship between savings and investment into proper perspective. Then maybe we'd be able to fix some of the macro models that are being passed around as having predictive power.

[Winslow R.]

James Kroeger wrote:

"You guys might be interested in some of my attacks on the presumed legitimacy of the SAVINGS = INVESTMENT identity. "

I've read some of your 'stuff' in the past and believe you are on the correct path. If you haven't read 'Understanding Modern Money, The Key to Full Employment and Price Stability', I'd recommend.

Sometimes it seems the people that are interested in these things numbers about 10 compared to the number of people that are affected - 6 billion. I am really glad Mr. Snow makes it 11. Too bad he may be losing his job.

[knzn]

Winslow: “What are 'real savings', nonfinancial assets?” Financial assets are not net savings, because every financial asset is also someone else’s liability: someone is dissaving just enough to offset the gross savings.

[Winslow R.]

Winslow: “What are 'real savings', nonfinancial assets?” Financial assets are not net savings, because every financial asset is also someone else’s liability: someone is dissaving just enough to offset the gross savings.

Two points:

First, you wrote The point is, prices change, but the total quantity of real savings does not. By your own definition, the quantity of 'real savings' increases with increases in nonfinancial assets. I agree this is how nonfinancial savings increase.

Second, I agree most financial assets net to zero as each credit has a debit. But I see a 'special class' of financial assets created by deficit spending, the issuer of the 'unit of account' that are only offset by future tax liabilities which may/may not be collected.

Why is this important? Credit/Debit expansion is limited by reserves and currently more so by capital requirements.

http://en.wikipedia.org/wiki/Capital_requirements.

The Capital tends to be made up of the type of financial assets I mentioned above, those created by deficit spending. There is a limit to the credit/debit expansion based on a country's income structure, its willingness to extend term lengths, its willingness to drop interest rates, and its Capital requirements.

I would argue there are nonzero financial savings that are equal to the U.S. government deficit. These financial savings have replaced gold as the reserves at the world's CB's. These financial assets have no 'liability' except possible future tax liabilities. They may even be bettter than gold because they earn interest but then again, if there are too many, they may not.