The Current Account in the Macroeconomic Adjustment Process

Questions & answers

Q1. What are the paper’s two central analytical claims?

The paper’s thesis is that (1) “an econo[m]y’s current account balance depends as much on future economic trends as on the current economic environment,” so a policy shift “will have one effect on the current account if it is perceived to be temporary and another if it is seen to be permanent,” and (2) “temporary disturbances in the econom[y] have permanent effects, by altering the entire future path of the economy’s international indebtedness” (Abstract). The paper develops both claims through a single formal model rather than treating them as separate results: the first follows from writing the current account as a function of the gap between current and permanent (annuity-equivalent) income, spending, and demand; the second follows because a current account deficit run today is a change in the stock of net foreign assets that persists forever once the shock itself has passed.

Q2. What are the paper’s organizing assumptions, and how does the model treat government financing?

The model assumes continuous market clearing and full employment throughout – a deliberately classical framework – although the author states that “the importance of the intertemporal dimension in current account determination will remain unchanged” even under Keynesian, non-market-clearing conditions (Introduction). Households are infinitely lived, intertemporally optimizing, and face a constant world real interest rate r*; with these assumptions (and non-distorting taxes) a Ricardian-equivalence result holds, so “consumer expenditure, the trade balance and current account will be invariant to the path of taxes and government deficits, given the path of government expenditure” – only the time profile of government spending, not how it is financed, affects the current account (Sec. 1, citing Barro 1974).

Q3. How is the economy’s production and consumption side set up?

Output is produced in two sectors – a purely traded good (the numeraire, in perfectly elastic world supply) and a semi-tradeable good facing a downward-sloping export demand schedule shifted by an exogenous foreign-demand variable – with an exogenous productivity shifter and a fixed domestic labor supply determining the production-possibility frontier between the two goods. Households maximize intertemporal utility over consumption of both goods with an additively separable, log-linear (“intertemporally homothetic”) utility function, chosen because it makes total consumption expenditure a linear function of wealth with a constant of proportionality equal to the discount rate – a tractability device the paper flags explicitly as not fully general (footnote 7: only the log form keeps that proportionality invariant to the future path of interest rates, though consumption remains linear in wealth for the broader CRRA class).

Q4. What is the paper’s key device for separating “temporary” from “permanent,” and how does it enter the current-account equation?

The paper defines, for any variable X(t), its “permanent” or “perpetuity-equivalent” value X^P(t) as the constant annuity (discounted at r) that has the same present value as the variable’s actual future path – i.e. X^P(t) solves ∫ e^{-r(T-t)}X^P(t)dT = ∫ e^{-r*(T-t)}X(T)dT.** Using this operator, the model’s solved-out current account equation (Sec. 1, eq. 15) decomposes CA(0) into: a time-preference term (r*-δ)W(0), positive whenever the world interest rate r* exceeds the household’s pure rate of time preference δ; a consumption-smoothing term in the deviation of current foreign demand from its own permanent level, (1-α)[W*(0)-W*^P(0)]; and a fiscal term in the deviation of current government spending from its permanent level. The paper states this decomposition shows three sources of current-account deficits, “all related to household preferences for certain consumption paths and to the intertemporal budget constraint”: low current relative to permanent income (households dissave to smooth consumption), a wedge between the world return and the discount rate, and government spending above its permanent level (Sec. 1).

Q5. What do the comparative-dynamics exercises show about temporary versus permanent versus anticipated shocks?

The paper works through three stylized perturbations to government spending – a temporary shock lasting until date T, a permanent shock, and a shock announced in advance of its actual onset – and shows each produces a different current-account and relative-price path. A temporary rise in government spending lowers permanent wealth (because some of the extra spending must eventually be paid for out of future absorption), running a current-account deficit while the shock lasts; critically, “the temporary fiscal expansion unambiguously reduces [wealth]… [and] raises Q^T(t), for t > T,” so even after the shock itself has ended, the now-lower stock of wealth leaves the relative price of the semi-tradeable good and its output permanently lower than they would otherwise have been – the paper’s central “temporary shock, permanent effect” mechanism made concrete (Sec. “A Formal Model,” comparative dynamics). For a pre-announced future cut in government spending (explicitly likened to a “Reagan-type” fiscal announcement), the model shows wealth rises immediately on the announcement (since permanent spending falls right away even though actual spending has not yet fallen), causing relative prices to move toward the semi-tradeable good and the current account to deteriorate before the policy takes effect, with a reversal in direction once the cut is actually implemented; for sufficiently long pre-announcement horizons T the author concludes “the long-run effect is almost surely a depreciation” of the relative price of the home (semi-tradeable) good.

Q6. What role does capital mobility play, and how does the paper show this using a no-capital-mobility benchmark?

The paper explicitly contrasts its baseline (perfect capital mobility, current account free to move) with a counterfactual in which the current account is constrained to zero every period, and shows the “temporary shock has permanent effect” result is specifically a product of capital mobility. Under the zero-current-account constraint, “temporary shocks during t < T have no effect on resource allocation for t > T” because households cannot borrow or lend to smooth consumption across the shock window – any attempt to do so is choked off by domestic interest-rate adjustment until the current account is balanced each period. The paper notes the direction of relative-price effects also differs qualitatively between the two regimes (governed by λ - α rather than λ - α(1-e^{-r*T})), underlining that both the existence and the sign of the long-run effect depend on whether intertemporal borrowing against the rest of the world is available.

Q7. What does the paper identify as the most important limitation of its baseline model, and how does it address this?

The author states explicitly that “perhaps the most important deficiency of the simple model is the absence of investment in physical capital,” since shifts in investment opportunities “give rise to strong current account effects in theoretical models” and, citing his own earlier empirical work, “seems to confirm the strong, even dominant, role of investment in cross-country current account behavior” (Concluding section). The paper does not formally extend its analytical model to include investment – it states this “enormously complicates the algebra, and typically forces a retreat to simulations” – but reports, citing companion simulation work, that permanent increases in world demand tend to produce initial current-account deficits (as capital formation draws in foreign financing) and that resource-allocation effects of temporary shocks are magnified, not dampened, once capital accumulation is allowed, because long-run sectoral supply elasticities rise.

Q8. What other extensions does the paper flag as capable of changing its conclusions, and how does the paper qualify its own generality?

The paper explicitly flags two further simplifications it has not relaxed within the model itself: a time-varying rate of time preference (the Uzawa formulation in which the discount rate depends on the level of utility, following Obstfeld 1980/1981) can change whether a given shock produces a surplus or a deficit, and replacing the infinite-horizon household with a finite-planning-horizon household “has even stronger effects on some of our conclusions,” since in that case government tax and debt policy – not just the spending path – can affect the current account and resource allocation through intergenerational redistribution (Concluding section, citing Buiter 1981 on an overlapping-generations variant). The paper is therefore explicit that its headline results – Ricardian irrelevance of financing, and the specific size/sign of long-run relative-price effects – are conditional on the infinite-horizon, constant-discount-rate, log-utility structure chosen for tractability, even while its author argues the core qualitative principle (forward-looking current-account determination; temporary shocks with permanent effects via the asset stock) is “certainly robust to changes in model specification.”