Read the [...] interview with [Michael Rivero] dated 11.22.2002. [...] Focus ... on what he says about USA debt and its overwhelming magnitude being a motivator for the on-going invasion of the Middle East. Basically he argues that acquiring the oil by force is a means to help pay off the USA's debt to certain preferred lenders. If true, it is certainly not generating a very fast payback for cash invested. So: not sure I agree with him or not, but would like to hear how your model of economics responds.

First, readers are gently requested to suspend whatever judgments they may have about Mr. Rivero's overriding opinions. In my private reply, I mentioned that he came off poorly. What I don't think I made adequately clear is that theory itself is worth examining, even if there are some rather extravant elements to his specific claims.^1.1 However, I was interested in the problem of income transfer. The "Transfer Problem" is the analytical question of how a unilateral transfer of financial assets, such as aid or reparations, affects the balance of payments of the affected countries.

At first, this seems like a simple matter: the USA gives $1 billion to the tsunami-affected countries, and their current account balance is shifted favorably by a billion, while that of the USA is shifted negatively by the same amount. The matter is made more complex by the fact that the countries will spend this (over infinite horizons) on increased purchases of goods and services, chiefly from the USA. This, in part, because our aid included conditions to that effect, but also because a transfer of US dollars with or without strings attached means an overall increase in claims against the USA that can only be met by US real output.

In 1929, John M Keynes and Bertil Ohlin carried out a debate over the "transfer problem" with some startling effects. The debate was stimulated by Keynes' Economic Consequences of the Peace (1920), which included some analysis of the trade effects of the Peace of Paris. The debate was over the transfer would be "overeffected," "effected," or "undereffected," that is to say, whether the effect on the current account balance would be net favorable, neutral, or net negative.^1.2

Ohlin had introduced the idea of a multiplier; this was later adopted by JMK, who made it the core of his General Theory. The increased income was accompanied by a marginal propensity to spend income on imports, which was (and is) supposed to be bigger than spending on imports as a share of GDP. Hence, an increase in income would lead to an increase in imports, while the transferor would experience a reduction in spending on imports.

Decades passed, the General Theory was published in 1936, then John R Hicks introduced the IS-LM chart, and then came two groups of researchers. One, Robert Mundell and Marcus Fleming introduced the Mundell-Fleming treatment of open economy macroeconomics (the IS-LM model in which there is trade and capital flows). This was pursued with fixed and floating exchange rates, leading to some startling conclusions. The other pair of researchers were S. Laursen and L.A. Metzler, who (1950) published what would become the standard analysis of the transfer problem for about 45 years.

In this model, L&M proposed a balance curve BB, which was simply all the various combinations of economic output and exchange rates for a developed economy that lead to a balance of trade. In standard usage, r represents the price of foreign currency in terms of the domestic currency. Hence, if r_USD should increase, then the USD has depreciated. Under this model, analysis focuses on two equations:

Movements of the BB curve up or down are believed to represent the effects of a transfer. If the transfer occurs regularly (as, for example, the effect on the US economy of net factor income), and if it is positive, then the BB will be higher. Ohlin had concluded that the transferor nation would experience a loss of imports equal to µT, where µ was the marginal propensity to spend on imports; the other country would increase imports by µ*T (where µ* was the transferee country's marginal propensity to consume imports); and BB would be higher at each value of i by an amount T - (µT + µ*T), where (µ + µ*) were probably much less than one.

The same effect would be noticed with RR: at points below RR, aggregate demand would exceed national income. RR would now increase, since income had increase. For each level of r, the level of national income at which A + B = Y would now be higher. A sustainable equilibrium would be where the two curves intersect.

However, some of my readers will notice a caveat: the upward slope of BB. The underlying assumption is that, at high levels of r, our terms of trade may be lousy, but our balance of trade will be stellar: cheap exports, costly imports—that will surely sustain balanced trade at high levels of Y, no?

Maybe not. In 2000, P. Hooper calculated price elasticities for the G-7 countries and arrived at a combined import-export price elasticity of -1.8 for the USA, suggesting that a 1% increase in the trade-weighted price of the dollar would increase imports by about 0.3%, and reduce exports by 1.5%.^1.3 The USA's exports were especially sensitive to long-run price changes. So my own regression results, which suggested a J-curve are apparently not so bizarre.^1.4 A J-curve would mean that the initial effects of a currency shift would lead to import-export price elasticites summing to less than 1; "initial" could also means "small." Depending on where one is on the curve, the effects on a rise in exchange rates on the trade balance could be positive or negative, a condition I have tried to illustrate here:

Unemployment statistics suggest the USA is at close to "full employment," yet this hides severe sectional dislocations; additionally, we are manifestly at a position above our BB curve. If the USA is the recipient of a net transfer of oil rents from Iraq, we would be closer to the BB curve than we were before production in that country was restored. On the other hand, the hoped-for rents might still require a devaluation of the USD to achieve overall balance anyway.

AFTERTHOUGHT: The fact that the Iraq War is costing more than the rents does not affect the figures, since the money is mostly spent on US-manufacted goods and services.

For information on "the transfer problem" I relied heavily on International Finace and Open Economy Macro-Economics, by Giancarlo Gandolfo, Springer-Verlag, 2002, p.110-127

The "transfer problem" unites these obsessions. One of the goals of the invasion of Iraq was the control of oil rents (streams of revenues) for certain elites. By itself, I doubt this would have motivated the invasion, and I don't believe the rentseekers back of the late invasion of Iraq were at all interested in distributing the rents nationally; I think rents from the US Treasury were, if anything, an even bigger prize than those plundered from Iraq. Certainly, there have been at least $9 billion in US reconstruction aid transferred from taxpayers to persons unknown and unknowable; this must be added to the billions in oil revenues seized or "vested" by the CPA.^2.1 Now, I want to make some points here: one is that the recipients of the vested assets of Iraq are obviously not planning to reward the taxpayers who paid for this monstrous adventure; if they did, the taxpayers would still be out a cool grand apiece. The war would not only look silly, it would have been silly. Having said as much, there is still a transfer problem to be discussed, which I've done here.

The gist of my analysis was that, while the USA was pouring "billions into Iraq" figuratively, in a national income & product accounting sense this was not necessarily so: the billions were being spent on American goods and wages, even if they were being shipped to Iraq for ultimate wastage. The hidden transfer of Iraqi assets to "US" expropriators would be felt, according to a conventional analysis, as an improvement in the terms of trade and in the equilibrium balance of payments....or, perhaps not.

The phenomenon of transfers affecting the terms of trade and balance of payments is known as the Harberger-Laursen-Metzler effect (HLME).^2.2 It was the first problem of open economy macroeconomics, and a decade later spawned the extremely important Mundell-Fleming architecture of macroeconomics. Keynesianism was to become a piece of the Mundell-Fleming puzzle, a subroutine of the far more comprehensive MF framework for understanding international trade.

The HLME is a flexible model, as I believe I implied in my previous essay on it. For example, the model allows one to consider a regime under which currencies float, rather than remain pegged. Also, one can examine the possibility that the exchange rate has a nonlinear relationship to the trade balance. Finally, and I regret not making this more clear earlier, the HLME is different from the classical supply/demand functions insofar as we don't assume, or even expect, the economy to lie on the designated equilibrium point. Like the IS-LM function, the diagram is really a map suggesting the lay of the land. If one lies on a point above the intersection of the RR-BB curves, one will experience excessive demand for goods and a trade deficit. In the chart below, diagram (a) represents the way the RR and BB lines have been presented in most textbooks.

In area β, there is underemployment in the domestic economy, but there's also a trade deficit. Devaluating the local currency may solve both problems, since inadequate domestic demand might be offset by increasing exports. Area γ is the converse of β, while δ is the converse of α.^3.3

In diagram (b), the situation is made slightly more complex by the fact that the elasticity of imports and exports is nonlinear; the extreme ends of BB should be ignored, since the curve is naturally a trend line drawn from observed history. Area β describes a situation impossible in (a), where the economy is running a trade surplus but there is inflationary overemployment; area γ likewise describes a condition where there is a domestic recession and a trade deficit. If the USA is on chart (a), then we could be in area β; if on chart (b), then we're in area γ. The difference is significant: a depreciation of the USD has so far accompanied a worsening of the US trade balance, while studies by Obstfeld & Rogoff (PDF; HC discussion) believe a continued increase of r would reverse this. On the other hand, it would also mean a chronic recession and possibly a permanent condition of massive current account deficits—even with a trade surplus. Since this tends to conform to societies I've read about, I think this latter analysis is the more accurate scenario.

2 For a brief professional introduction to the Harberger-Laursen-Metzler effect, see "the Harberger-Laursen-Metzler effect Revisited" (PDF), Roberto Duncan, Banco de Chile, 2003; otherwise, check the index of any "Open Economy Macro" textbook.

In a couple of entries I've speculated on the impact of a transfer of income from Iraq to the USA as a result of corrupt CPA management (1, 2). As I have been very anxious to make clear, I do not believe the invasion of Iraq was affected in order to achieve a transfer, and in any event the transfer affected will be negligible relative to the US current account deficit. Even the mere abolition of the subsidies to SUV buyers would have a more prominent effect on the trade balance (SUVs themselves are mostly manufactured in the EU).

However, the transfer effect is still of considerable interest. The US economy is already mightily afflicted by reccuring transfers caused by a huge trade deficit and concomitant accrual of foreign-owned, income-earning assets. Should this lead to a deterioration of American terms of trade with the rest of the world, there's a speculation—the Harberger-Laursen-Metzler hypothesis—would led to a reduction in the real income, and with it, a reduction in domestic saving. This would require an even greater demand for net capital investment from abroad. Put another way, the effect of a trade deficit could be a decline in the purchasing power of the US dollar...leading to a worsening of the trade deficit.

Curiously, this is one of the older propositions in Keynesian economics, seldom challenged, but also seldom invoked. It is obscure and counter-intuitive. In 1982, the economist Maurice Obstfeld challenged the plausibility of the HMLE effect; he argued that a permanent terms-of-trade deterioration would cause real expenditures to decrease sharply, and thereby improve the current account.^3.1 Others refined this to examine specialized scenarios, such as if the economy belonged to a small country or a large one, or if the items being imported tended to be luxuries... or the duration of the trade shocks.

As readers can probably tell, I've been skimming through recent literature on the HLME to help reach a conclusion about the effects of trade shocks such as the one our economy is in right now. Most of the literature is based on theoretical models in which households optimize utility functions, based on increasingly specific assumptions. Among the flurry of competing models is an actual study of the historical evidence by Abdur R. Chowdhury, which includes a regression of domestic private saving on many components, including terms of trade shocks.^3.2 Disappointingly for my purposes, the study concentrates on small economies (all in Central Europe). Still, the regression results are a wealth of insights. Please be mindful that Chowdhury's dependent variable is private saving, which tells us how domestic demand for goods (including imports) responds to various factors:

The coefficient on the lagged private savings rate is, as expected, positive. The value of 0.529 shows the presence of a large degree of persistence. [...] Countries with higher per capita income tend to save relatively more than countries with lower per capita income. [...]

The financial depth variable (measured by the ratio of M2 to GDP) has a highly significant negative impact on private savings. When the volume of M2 rises by 1 percent of GDP, the private savings rate decreases by 0.24 percentage point. This result confirms the widely held view that financial reforms may stimulate consumption by relaxing domestic liquidity constraints e.g. through increased access to bank credit, and thus reduce the propensity to save. [...] Inflation has a positive impact on savings. An increase of inflation by 10 percentage points raises private savings by eight-tenths of one percentage point. The results indicate that increased uncertainty about the aggregate economy and expectations of further price increases induce agents to lower their current consumption and increase precautionary savings.

The coefficient on public savings is negative and statistically significant, which suggests that the private sector internalize the government’s budget constraint. The short term coefficient is 0.176 giving a permanent long-term value of 0.373. Since the coefficient is statistically less than one, we can reject Ricardian equivalence for the full sample.
[p.19]

Both the permanent and temporary components of the terms of trade are positive and statistically significant. [...] Moreover, the magnitude of the coefficient on the temporary component is much larger than that of the permanent component. This reflects the lack of access to foreign borrowing that many transition economies faced during the 1990s.* The short-term coefficient on the transitory variable is 0.274, so the long-term effect is around 0.582. As both values are significantly less than one, there is an incomplete pass-through in the system. This may be due to the inability of the households to fully judge the persistence of the terms of trade shock at the moment it occurs.
[p.20; * refers to a liquidity constraint that the USA has heretofore avoided.]