Questions Raised by the Eurozone Experience

Adoption of the euro as a common currency altered the macroeconomics of member states by officially removing the possibility of nominal exchange rate devaluation as a mechanism of relative macroeconomic adjustment. Flexible exchange rates as a mechanism of relative price adjustment has been an important element in economic theories of Freidman (1953) and Keynes alike. The need for adjustment in international relative prices and mechanisms for facilitating it remains an active research area, and the macroeconomic experience of the eurozone has provided new evidence and raised new questions for researchers.

To provide a partial overview of the macroeconomic experience of the eurozone, Figure 1 plots the growth rate in gross domestic product for selected member countries. It presents data from Germany as representative of the eurozone core and three countries located on the eurozone periphery: Spain, Italy, and Greece. In theory, countries suited for a currency union should have similar business cycle fluctuations, given that they do not have recourse to independent monetary or exchange rate policies. The figure shows that some recession episodes, notably the Global Financial Crisis of 2007–2009, was shared broadly among European countries. But other recessions, notably the eurozone crisis of 2010–2012, was highly asymmetric among countries. In particular, Greece experienced a severe recession starting in 2010, as large government debt levels were revealed, which prompted the imposition of fiscal austerity. Other periphery countries such as Spain and Italy also went into recession the following year. The macroeconomic experience was strikingly different in Germany (also in France, not shown in the figure), which did not officially enter into recession. Clearly, the co-movement of business cycles among these countries depends upon the nature of the shock.

Open in new tab

Figure 1. GDP growth (quarterly, year over year) for selected eurozone countries.

Source: International Financial Statistics, IMF.

Figure 2 brings focus to this question by plotting real exchange rates since the introduction of the euro. The usual definition of the real exchange rate is eP^*/P, where e is the nominal exchange rate in units of foreign currency per home, P is the domestic consumer price index, and P^* the foreign price index. Given that the nominal exchange rate between any two members of the eurozone is defined to be unity, the real exchange rate here is the consumer price index of a country relative to that of a base country (Germany, in this case). The fact is, the introduction of the currency union did not lead to a convergence to a common price level, and inflation rate among member countries posed a research question in the early euro period: How can prices differ if there is a single currency? The figure shows rising relative price levels in all periphery countries relative to Germany in the first decade of the euro. These reached a peak at the time of the eurozone crisis and have backtracked to their initial levels since. Greece stands out in this regard, in that its relative price had risen 20% by the time of the crisis, and since it has abated more than halfway back to its pre-accession level.

Open in new tab

Figure 2. Relative CPIs (relative to Germany) for selected eurozone countries (base year of 1.00 in year of eurozone accession).

Source: International Financial Statistics, IMF, and author calculations.

This swing in real exchange rates is reflected in current account balances (net exports in goods, services, and income), plotted in Figure 3 as ratios to gross national income. During the period that real exchange rates were appreciating, countries lost competitiveness and had negative net exports. This was partly reversed after the onset of the crisis, which alternately could be attributed to the improved price competitiveness as real exchange rates depreciated or to the drop in imports and absorption during the recession.

Open in new tab

Figure 3. Current account as a share of gross national income.

Source: International Financial Statistics, IMF, and author calculations.

Early research proposed a narrative explaining these facts in terms of productivity levels in periphery countries catching up with the European core. From this perspective, the real exchange rate appreciation of periphery countries was a natural equilibrium process and even favorable. (See Blanchard & Giavazzi, 2002, for a defense of this more sanguine view.) However, an alternative explanation lies in stronger demand in periphery countries, perhaps arising from access to cheaper eurozone credit, which facilitated fiscal imbalances (see, e.g., Canova & Pappa, 2007). The difference in these two interpretations centers on whether the real exchange rate appreciation reflected fundamental equilibrium forces—or instead reflected unsustainable demand policies. The eventual resolution involved Greece (and to varying degrees other periphery countries) enduring a severe economic downturn to bring down domestic wages and prices and thereby depreciate the real exchange rate. The Greek experience inevitably raises a question of whether the adjustment process of the real exchange rate would have been less painful if Greece had retained an independent nominal exchange rate and could have utilized a nominal devaluation.

The macroeconomic experience of the eurozone points to a fundamental question in international macroeconomics regarding the need for adjustment in international relative prices and mechanisms for facilitating it. The next section will discuss empirical research regarding the role of the nominal exchange rate in facilitating macroeconomic adjustment, with special focus on the evidence for the European case. The third section will discuss theoretical research regarding when giving up the nominal exchange rate as a mechanism of relative price adjustment is costly in terms of welfare. The fourth section will discuss theories of alternative mechanisms for adjustment and compare them to nominal exchange rate devaluation. As a natural counterpart to reviewing current thinking as embodied in recent literature, this article will also highlight unresolved questions that would be productive topics for future research.

Empirical Evidence Regarding the Nominal Exchange Rate as Adjustment Mechanism

Empirical research has studied changes in the behavior of the real exchange rate after the introduction of the euro to determine how relative price adjustment has been affected by elimination of the nominal exchange rate as an adjustment mechanism. Some of this research has focused on the average level of real exchange rates as a measure of the magnitude of price misalignments across countries, while other research has focused on the volatility of the real exchange rate as a measure of how often shocks push relative prices away from their long-run equilibrium, and still other work has studied the speed at which the real exchange rate adjusts over time to return to its long-run equilibrium. One perspective on the literature is that there is still, at this stage, a lack of consensus regarding the net effect of the euro on the magnitude and persistence of price misalignments. Nonetheless, this literature has served to identify a set of key factors at work, as well as raising a set of key questions for future work to address.

Several studies that use data on absolute price levels of goods find the degree of price dispersion within eurozone countries was lower after they adopted the euro (see Fidora, Giordano, & Schmitz, 2017, as well as Glushenkova & Zachariadis, 2016, which builds on Crucini, Telmer, & Zachariadis, 2005). This finding would suggest increased integration of national goods markets after adoption of the common currency. But Berka, Devereux, and Engel (2012, 2018) noted that the average level of price dispersion among eurozone insiders is not distinguishable from that of European countries outside the eurozone. Instead, their work argues that the main distinction between the real exchange rates of the eurozone and the floating rate countries of Europe is to be seen in the differences in real exchange rate volatility (i.e., the tendency of the real exchange rate to deviate from its mean). They find that the standard deviation of the real exchange rate is much higher among European countries retaining flexible exchange rates compared to countries within the eurozone.

To investigate the source of this higher volatility, Berka et al. (2012, 2018) decomposed the real exchange rate into two parts: (1) the relative price of internationally traded goods across a given pair of countries; and (2) the relative prices of nontraded to traded goods within each country. This distinction has significance, given the shared theory of Balassa (1964) and Samuelson (1964), in which a rise in relative productivity of the tradable goods sector in a country leads to an appreciation of real exchange rate in that country. In this mechanism, higher productivity leads to higher wages, which induces a rise in the price of nontraded goods; since the price of traded goods is pinned to the world level by international arbitrage, the real exchange rate appreciation is attributed to the rise in relative price of nontraded goods to traded goods within a country.

The higher real exchange rate volatility of non-eurozone countries appears to be due to the traded goods component rather than the internal relative price of nontraded goods (Berka et al., 2012, 2018). Further, it is found that the real exchange rates in the eurozone countries are more highly correlated with the relative price of nontraded goods for European countries, which is consistent with the Balassa-Samuelson theory. As further evidence, the authors collect measures of the relative productivity of traded goods sectors and find that the real exchange rate is much more correlated with the measure of productivity for eurozone countries. This evidence suggests that Balassa-Samuelson dynamics played at least some role in driving real exchange rate developments in the eurozone, and they certainly played a greater role there than for flexible exchange rate countries outside the eurozone. Berka et al. (2018) strengthened this argument by controlling for unit labor cost in their regressions and showing that the remaining real exchange rate fluctuations are even more tightly linked to the relative price of nontraded goods. Their argument for doing this is that unit labor costs are largely the product of a separate set of economic and political forces distinct from productivity shocks and are best viewed as the result of shocks to labor market frictions (i.e., a labor market wedge).

This set of results has been used as evidence against the claim that sacrificing a nominal exchange rate by adopting the euro has harmed countries’ abilities to effect relative price adjustment. It further is argued in Berka et al. (2012) that the euro on balancehas been helpful in facilitating adjustment. Since movements in the relative price of nontraded goods can be tied to fundamental macroeconomic changes in productivity, the fact that real exchange rates can be tied to these internal relative prices in the eurozone indicates that the real exchange rate movements are serving a useful function of macroeconomic adjustment. Conversely, the fact that outside the eurozone real exchange rate movements are due mainly to deviations from the law of one price indicates that they are generating additional relative price distortions and misalignments that detract from economic efficiency, rather than providing useful macroeconomic adjustment.

One concern regarding this interpretation is that while Berka et al. (2012, 2018) showed that the Balassa-Samuelson mechanism played at least some role in driving real exchange rate movement in the eurozone, it is unclear how important they are in explaining the facts of the eurozone experience, especially during the eurozone crisis. Berka et al. (2012) is clear in stating the caveat that their data end in 2009 before the eurozone crisis. Nevertheless, it is left unclear what fraction of real exchange rate fluctuations, such as those depicted in Figure 2, are due to the relative price of nontraded goods. It would seem that a formal variance decomposition of the real exchange rate fluctuations would be a useful exercise in future work. If a large fraction of the real appreciation of countries such as Greece and Spain are due to rises in unit labor costs unassociated with productivity gains or the Balassa-Samuelson mechanism, then this leaves unexplained a major part of the euro macroeconomic experience.

Nonetheless, there seems to be broad consensus around the general point that when the nominal exchange rate is flexible, a large share of its movements are unrelated to macroeconomic fundamentals; and further, this type of nominal exchange rate fluctuation can be a source of disruption to international relative prices rather than a productive mechanism resolving macroeconomic imbalances. A currency union clearly alters this scenario by precluding nominal exchange rate fluctuations. The theoretical explanation typically offered for this set of empirical results has two key ingredients. First, nominal exchange rate movements are driven mainly by financial market shocks, unrelated to macroeconomic fundamentals. Global financial integration has expanded to the point that the large majority of demand for foreign exchange reflects asset market transactions rather than international trade in goods. Financial market shocks appear to dominate determination of the nominal exchange rate, leading to volatile fluctuations unrelated to the needs for equilibrium in international goods markets.

The second key ingredient is that prices tend to be sticky in terms of the local currency of the consumer, so that nominal exchange rate fluctuations do not affect the relative prices of imports that consumers see, and thus they cannot serve as a useful international macroeconomic adjustment mechanism. While Berka et al. (2012, 2018) appeal to this logic using a New Keynesian model, this general point was made, and even applied to the case of the incipient eurozone, by Engel (2000) using an older Mundell-Fleming model to argue the point that surrendering the nominal exchange rate as an adjustment mechanism may not be costly. Mukhin and Itskhoki (2019) showed that the same result can arise in a model with financial market segmentation instead of the standard sticky price New Keynesian model. Nominal exchange rate risk is held by a small group of financial intermediaries and not shared smoothly throughout the economy. Under this type of financial friction, one again sees that exchange rates are driven by financial market developments and are disconnected from fundamental forces of the goods market.

These conclusions, arising from the empirical literature studying real exchange rate volatility, are echoed and augmented in the related empirical literature studying the persistence properties of real exchange rates. This literature focuses on the speed at which the real exchange rate adjusts to correct deviations from a long-run equilibrium value. The concept of long-run equilibrium in this literature usually is taken from purchasing power parity (PPP), in which the overall price level should be the same across countries once adjusted to the same currency. Typically, real exchange rate changes are regressed on lagged levels of the real exchange rate deviation from the long-run equilibrium. The literature contains a wide range of estimates for the speed of adjustment. Many find evidence of stronger convergence to long-run equilibrium in samples that include the euro (Bergin, Glick, & Wu, 2017; Koedijk, Tims, & Van Dijk, 2004; Lopez & Papell, 2007), while some papers find the opposite (Huang & Yang, 2015), and one interesting finding suggests slower convergence after the euro but an acceleration after the onset of the euro crisis (Fidora et al., 2017). The fact that some papers find faster convergence after the introduction of the euro may seem surprising, given the loss of the nominal exchange rate as a mechanism to compensate for sticky goods prices. But in the context of disruptive exchange rate fluctuations unrelated to goods prices, this finding can be interpreted as consistent with the result of Berka et al. (2012) that adoption of the euro need not hamper relative price adjustment.

Particular evidence in support of this claim is found in studies that decompose the real exchange rate into the nominal exchange rate and the ratio of average price levels in each country’s own currency (see Bergin et al., 2017). Estimates of the dynamic response of each of these components indicate that the volatile shocks to the nominal exchange rate arising from financial markets tend also to be highly persistent, so the fact that the euro eliminated this source of real exchange rate fluctuations implies a faster average rate of adjustment in real exchange rates to their long-run equilibrium. Again, it appears that elimination of the nominal exchange rate eliminated a source of shocks and misalignment even more than it eliminated a useful mechanism of equilibrium adjustment.

This research also found a second source of faster real exchange rate adjustment during the euro period: National price levels responded more strongly to correct the disequilibrium. It appears that once the euro eliminated the nominal exchange rate as a mechanism of adjustment, prices became somewhat more flexible to take on more of that role. This potentially might result from greater price transparency, given that prices in different countries were denominated in a common currency, promoting stronger forces of arbitrage and price adjustment. Future work is needed to better understand the mechanism for this result, perhaps by better integrating the price dispersion from microeconomic evidence with the time series work, or by integrating the distinction between traded and nontraded goods, and redefining the concept of long-run equilibrium in terms of the Balassa-Samuelson theory rather than PPP.

Some evidence points to an evolution over the lifetime of the eurozone, indicating that the speed of real exchange rate adjustment became faster after the eurozone crisis. This may be evidence that prices have become more flexible over time in the euro. The result may well be driven by the downward adjustment in real exchange rates observed in Figure 2 for countries such as Greece in the post-crisis period. One question unresolved by this finding is that even if adjustment can occur despite the absence of a nominal exchange rate, adjustment via goods prices might be a more painful mechanism, in terms of output or welfare foregone, than adjustment via the nominal exchange rate.

Hints at an answer to this question can be found in some work predating the euro. Parsley and Popper (2001) estimated speeds of real exchange rate convergence in a panel of countries including but not limited to Europe in a sample ending prior to the euro but including the Bretton Woods system where many countries fixed their exchange rate values relative to the U.S. dollar. In addition to finding faster convergence under a dollar peg, the paper also found a prominent nonlinear effect, in which larger real exchange rate deviations from long-run equilibrium elicited faster convergence, and this effect was particularly strong for a dollar peg. One is tempted to interpret this result in light of the experience of the euro crisis. Under conventional exchange rate pegs (e.g., Bretton Woods or the European Exchange Rate Mechanism that preceded the euro), there were mechanisms for countries to alter the value of the peg or the central parity if the pegged value was too far from an equilibrium value consistent with international balance. These regimes reduced the noisy exchange rate fluctuations plaguing a flexible regime but still permitted infrequent exchange rate adjustment when imbalances were most severe. Sometimes this was messy, as in the case of speculative attacks. But there were occasions in which European periphery countries during the period of fixed but adjustable exchange rates devalued their currency pegs, as a means of releasing pressure building up from spells of higher inflation and hence appreciation in real exchange rates. Such a use of devaluation is consistent with the empirical evidence for past pegs in Parsley and Popper’s data set. But such devaluations are precluded by the irrevocable peg represented by the euro.

One can imagine that such a devaluation could have been an appealing option to Greece as a way of releasing pressure during the eurozone crisis and an easier mechanism of adjustment compared to the slow and painful process of lowering domestic wages and prices. So while there is empirical evidence that losing a nominal exchange rate (as under the euro) does not need to slow macroeconomic adjustment on average, there are times and scenarios, such as during the euro crisis, in which the loss of a flexible nominal exchange rate can be more costly.

When is the Loss of Exchange Rate Flexibility Costly?

There is a long theoretical tradition undergirding the idea that nominal exchange rate flexibility could compensate for the lack of price flexibility in promoting macroeconomic adjustment and that losing this mechanism is a key cost of monetary union. Keynesian models rooted in the Mundell-Fleming tradition have been used to articulate a “trilemma,” in which a policy of exchange rate pegging makes it impossible both to enjoy the benefits of international capital mobility and to use independent monetary policy to pursue goals of domestic macroeconomic stabilization. Similarly, the classic Mundellian approach to optimal currency areas (Mundell, 1961) characterizes the tradeoff between the loss of monetary policy independence and the benefits of lower trade costs for countries that trade extensively. Frankel and Rose (1998) has revisited both the benefit and cost sides of Mundell’s tradeoff to recognize that the criterion of trade integration is actually endogenous, so that entry into a monetary union can lead to greater trade volume. Further, it has been pointed out that for a country that lacks internal monetary policy discipline and credibility, the loss of monetary policy independence may actually be a blessing of a commitment to monetary stability (see Alesina & Barro, 2002).

A variety of models have been used to explain the benefits of an independent monetary and exchange rate policy for macroeconomic stabilization. In the case of a traditional Keynesian closed economy model, monetary expansion traditionally has been thought to work by lowering domestic interest rates to stimulate domestic investment and consumption demand. However, for an economy that is open in terms of global asset and goods markets, the channel for monetary policy tends to shift from the interest rate to the exchange rate, given that the interest rate is pinned down by asset market arbitrage relative to the world interest rate. If prices are sticky in the seller’s currency, a fall in the value of the domestic currency makes home exports cheaper for foreign purchasers, in principal raising domestic exports; likewise, it raises the price of foreign exports and lowers the level of domestic imports. This improvement in the trade balance provides additional demand for domestic production, though it may come at the cost of foreign producers, as a “beggar-thy-neighbor” policy. It bears remembering that one of the economic rationales offered for the creation of the eurozone was that nominal exchange rate devaluation in the past had been used within Europe to gain competitive advantage over trading partners. Given the elimination of national trade barrier protections within the European trading block, there was felt a need to preclude this type of competitive devaluation from taking advantage of the open borders. So, while exchange rate adjustment can be viewed as a valuable strategy for macroeconomic adjustment, it can be viewed also in some circumstances as a dubious one.

New Keynesian models, with more developed microeconomic foundations, have been used to quantify the welfare loss for European countries foregoing the exchange rate as a macroeconomic stabilization tool. Much of the qualitative story from a Mundell-Fleming model carries over to this environment, though Obstfeld and Rogoff (1995) were the first to note in such a model that expenditure switching leading to shifts in production across borders does not necessary result in beggar-thy-neighbor effects in terms of a formal welfare criterion based on household utility. In addition, Devereux and Engel (2003) noted that the expenditure switching effect disappears if prices are assumed sticky in the currency of the consumer rather than producer. In this case, movements in the exchange rate do not affect the relative prices observed by consumers and so do not induce changes in consumption decisions or expenditure switching.

The welfare effects of pegging have been found in this literature to depend on the source of exchange rate fluctuations. Kollmann (2002) studied an environment in which frictions in the international asset market allow for deviations from uncovered interest rate parity (UIP) that drive the nominal exchange rate. The optimal monetary policy rule implies targeting domestic producer inflation and involves substantial nominal exchange rate adjustment. Adopting a fixed exchange rate rule rather than the optimized rule stabilizing inflation implies a loss in welfare around 1% of annual steady state consumption. The peg requires that the domestic interest rate adjust more strongly to international financial market shocks, so these shocks have a more destabilizing effect on consumption. On the other hand, subsequent work (Kollmann, 2004) finds that adopting a monetary union potentially could raise welfare relative to a flexible exchange rate case. This occurs if the monetary union is assumed to eliminate the existence of the financial market shocks that otherwise would drive deviations from UIP and hence nominal exchange rate volatility. UIP shocks are harmful to welfare because they induce substantial reallocation of productive resources, in particular the use of imported intermediate inputs and labor.

However, subsequent work (Mukhin & Itskhoki, 2019) casts some doubt on this assumption that the adoption of a monetary policy rule fundamentally alters the exogenous shocks of the global financial market, suggesting instead that it simply alters the expression of these shocks on equilibrium outcomes. It remains a matter of interpretation and a subject of active research whether financial market shocks are best viewed as exogenous features of the financial market or are instead endogenous to the policy regime.

Determination of optimal monetary policy can be further complicated by a monetary union in which regions are heterogeneous. For example, if the degree of price stickiness differs by region, an analysis of optimal policy suggests that the inflation target used in formulating central bank policy should place greater weight on those regions with more price stickiness (Benigno, 2004).

The experience of the eurozone crisis has motivated the literature to study new types of costs of a peg. As highlighted in Schmitt-Grohe and Uribe (2016) many of the eurozone periphery countries experienced large capital inflows in the early euro period, which stimulated domestic demand, and induced a rise in average wage. But when the global financial crisis hit in 2007–2009, capital inflows stopped, and aggregate demand declined. Adjustment was hampered by the fact that wages remained at the high levels reached at the peak of the preceding boom, resulting in a large rise in unemployment. This circumstance underscored for many the cost of losing an independent monetary policy and the possibility of exchange rate devaluation by having joined the euro currency union.

The theoretical model offered by Schmitt-Grohe and Uribe (2016) suggested that the combination of fixed exchange rate along with international capital mobility and downward nominal wage rigidity creates a negative externality that leads to over-borrowing in good times and excess unemployment in downturns. The assumption of downward rigidity in nominal wage might be viewed as an alternative nominal rigidity to the stickiness in goods prices emphasized in the New Keynesian literature. The demand for labor from the nontraded sector depends upon the wage rate and demand for nontraded goods, which in turn depends upon the relative price of nontraded to traded goods and hence the exchange rate. Capital inflows lead to high demand for nontradables, which drives up the wage and puts the economy in a precarious situation, since in the contraction of the cycle, downward nominal wage rigidity prevents wages from falling to the level consistent with full employment. In the case of an adverse shock lowering demand, firms are not willing to lower prices to the level consistent with full employment, since production costs remain high due to the rigid wages.

Simulations of this mechanism show that the optimal monetary policy is inconsistent with a peg but would imply devaluations are needed to adjust the real wage to maintain full employment (see Schmitt-Grohe & Uribe, 2016 for details). The authors show that such an exchange rate policy in this environment is capable of achieving the Pareto optimal allocation: that is, eliminating the effect of the labor market distortion. A version of the model calibrated to the European crisis reveals that extremely large devaluations may be needed in practice to fully stabilize a case such as Greece.

This result differs in notable respects from the prior New Keynesian work. The assumption of sticky wages sidesteps the controversy in the New Keynesian literature over whether goods prices are sticky in the currency of the buyer or seller. The story also differs in that it does not imply a beggar-thy-neighbor outcome in which an exchange rate devaluation is designed to steal demand away from trading partners. This devaluation is not directed at generating expenditure switching to raise exports but rather aimed at remedying the friction in the labor market created by the downward nominal wage rigidity. The analysis also features an implication of financial market openness that is new. Unlike the models of Kollmann (2002, 2004), the focus is not on shocks arising in the global financial market (UIP shocks) but instead on an externality that arises from incentives to over-borrow during boom times. Such a story coincides with what was observed during the euro crisis for many periphery countries.

An important policy implication is that a case like the eurozone crisis might warrant limiting international capital mobility. Capital controls, such as taxes on international capital flows, then present a possible alternative set of policy tools in a monetary union. In the context of the model of downward nominal wage rigidity used to derive the optimal devaluation policy, one can also solve for an optimal capital control tax under the alternative assumption of a fixed exchange rate regime that prevents such a devaluation. Optimal policy restricts capital inflow during the booms and subsidizes it in bad times. Policy is thus prudential in nature, with the goal of preventing the run-up in wages during booms that becomes a problem during the slump. This research points to one type of policy alternative for a monetary union that has given up the nominal exchange rate as a policy tool—a topic that will be taken up as the subject of the next section in this article.

Alternative Adjustment Mechanisms for Europe

The preceding section recounted theoretical research showing potential costs to losing exchange rate devaluations as a tool of macroeconomic adjustment. This raises the question of what alternative policy tools are available to eurozone countries. Fiscal policy, of course, is a policy tool suggested by international macroeconomic theory dating back to Mundell-Fleming models; in fact, such models suggest that in the case of fixed exchange rates, fiscal policy should be a particularly potent tool of aggregate demand management. In a flexible exchange rate setting, a government spending increase or tax cut intended to raise aggregate demand may induce a rise in interest rate that crowds out domestic investment demand, or an exchange rate appreciation that crowds out export demand. But in the case where exchange rates are fixed and hence interest rates are pinned down, such a crowding out effect does not arise. Indeed, analysis of the optimal policy mix in New Keynesian models of a monetary union ascribe to fiscal policy the task of output stabilization, while the common monetary policy focuses on stabilizing overall inflation (Gali & Monacelli, 2008).

Unfortunately, in the particular case of the eurozone, recourse to the standard fiscal policy alternative is not readily available. Accession to the eurozone required that members commit to fiscal limits on debt and deficits as shares of GDP. The existence of such fiscal rules has been justified in terms of eliminating fiscal externalities biasing member countries toward larger debts, such as the temptation to shift the burden of unsustainable debts to fellow members, as well as the resulting excess inflation if a common monetary authority with imperfect commitment tries to mitigate potential debt crises caused by a member’s default (see, e.g., Aguiar, Amador, Farhi, & Gopinath, 2015; Beetsma & Uhlig, 1999; Bergin, 2000; Chari & Kehoe, 2007). Such fiscal rules have limited the tools available to eurozone periphery countries in dealing with their downturn during the eurozone crisis. In exchange for outside financial assistance Greece and others were required to accept a policy of fiscal austerity, the opposite of the usual prescription of a fiscal expansion.

Of particular interest in policy circles has been a nonstandard type of fiscal policy, termed a “fiscal devaluation.” The fundamental idea is to use combinations of distortionary taxes and subsidies to alter relative prices to give a country’s exports price competitiveness, so as to stimulate export demand and raise the trade balance. A basic version of this idea goes back to Keynes, who proposed the use of tariffs on imports and subsidies on exports to mimic the effects of a currency devaluation under the gold standard. The more recent incarnation consists of a revenue-neutral shift from employers’ social contributions toward a value-added tax. A reduction in the social security contributions lowers unit labor costs and hence producer prices for all firms, including those that export. The value-added tax raises the cost of consumption but does not impact exported goods; so in addition to balancing the revenue implication of the cut in the abovementioned social security taxes, it raises the price of consuming imported goods.

While fiscal devaluation has been a topic of discussion in policy circles for some time, Farhi, Gopinath, and Itskhoki (2014) provided a formal theoretical analysis in a New Keynesian environment (Schmitt-Grohe & Uribe, 2016; discussed in the section “When is the Loss of Exchange Rate Flexibility Costly?”) also included a simple treatment of a policy that resembles a fiscal devaluation. Among other things, this formal treatment indicates under what circumstances the tariff and VAT combination can fully replicate the effects of a currency devaluation and when they need to be supplemented with additional policy adjustments. It has been found that for a general environment, the standard fiscal devaluation (payroll and VAT tax changes) needs to be combined with a uniform reduction in consumption taxes and an increase in income taxes, in order to fully replicate all implications of an exchange rate devaluation. But under certain more restrictive conditions, such as complete asset markets between countries, and when the devaluation policy is unanticipated by the agents in the economies, changes in consumption and income taxes can be dispensed with—and the basic version of fiscal devaluation suffices.

The underlying complication is that the standard fiscal devaluation results in an appreciated real exchange rate relative to a nominal devaluation. Although the fiscal devaluation has the same effect on the terms of trade, it leads to an increase in the relative price of the home consumption bundle, which is not the case under nominal devaluation. This fact will not affect the real allocation when trade is balanced or when the devaluation is unexpected and asset markets are incomplete, since risk-sharing and household-saving decisions are not affected under these circumstances. By contrast, in the case of an expected devaluation, there will be an effect on savings and portfolio decisions, which would need to be undone with a reduction in consumption taxes: An offsetting increase in income taxes is required so as not to distort the labor supply decision of households. The situation becomes even more complex when the asset market takes the form of non-contingent bonds that are denominated in the home currency. In this case even the fiscal devaluation augmented with consumption and income taxes does not replicate the nominal devaluation, and equivalence then requires a partial default by the home country. (See Farhi et al., 2014 for a detailed discussion of these points.)

In addition, policymakers also point to additional likely limits to the applicability of a fiscal devaluation (see International Monetary Fund [IMF], 2011). For example, VAT rates in many countries differ among goods so that changes could induce distortions favoring some goods. And if non-labor income is not taxed, there may be additional distortions introduced to the labor market. It is important to note that this mechanism only works in an environment with a pegged exchange rate and rigid wages, otherwise the rise in demand for exports and reduced demand for imports will generate pressure in the foreign exchange market for a nominal exchange rate appreciation, which would undo the competitiveness impact of the tax shift. Likewise, if the wage rate were not rigid, there would be pressure for it to rise as workers find that the increased VAT tax rate reduces their real wage.

The limited empirical evidence available suggests that fiscal devaluation does have an effect on trade balances, but the magnitudes are moderate, and large fiscal devaluations would be needed to induce the necessary adjustment in eurozone countries. Most of this evidence comes from work with model simulations, which tends to suggest that a fiscal devaluation of 1% of GDP can raise the trade balance in the short term by only 0.1–0.6% of GDP and vanishes in the longer run (see Engler, Ganelli, Tervala, & Voigts, 2017). Econometric studies as yet are rare. Work by de Mooij and Keen (2013) uses a panel of 30 OECD countries and regresses the change in trade balance on measures of the social security and VAT tax. In their main empirical approach, using measures of tax revenue, a budget-neutral fiscal devaluation amounting to 1% of GDP improves the trade balance of a devaluing country by 2.8% for countries outside the eurozone, and finds a larger effect for the eurozone of 4%. In an alternative empirical specification using tax rates, the magnitude of the impact is lower and is statistically significant only for the eurozone countries. More work is needed on this important topic, with varying data sets and approaches.

When one considers the fiscal policies actually pursed by euro-periphery countries in the aftermath of the crisis, it is dominated by fiscal austerity, involving cuts in government spending and tax increases. Greece and Spain cut spending by 20–50% in the period 2010 to 2014 compared to 2009 and raised tax rates by 4–6 (Lambertini & Proebsting, 2019). This clearly is directly opposite to the usual fiscal policy expansion suggested by traditional Keynesian models as an alternative to an exchange rate devaluation. This fiscal austerity has sometimes been referred to as an “internal devaluation,” in that it was associated with a decline in the real exchange% rate (as seen in Figure 2), and sometimes it is interpreted as a variant of a fiscal devaluation policy. However, it probably is best viewed as entirely distinct. Notably, the fiscal devaluation policy cocktails are usually formulated to be fiscally neutral, whereas the recent austerity was specifically designed to reduce fiscal deficits.

Attempts to model this internal devaluation in New Keynesian open economy models (see Lambertini & Proebsting, 2019) indicate that while it generally leads to a depreciation of the real exchange rate, it by no means replicates all aspects of a nominal devaluation. The real exchange rate devaluation in periphery countries was associated with a fall in domestic wages, but this did not lead to a fall in a country’s export prices or expenditure switching in export markets; instead, the current account improved mainly because of the fall in imports by domestic consumers. Another clear contrast is that internal devaluations have been notably contractionary on output and employment in the short run. One might well argue that austerity could be beneficial in terms of macroeconomic adjustment in the long run to a sustainable long-run equilibrium, but it is hard to justify it as a substitute for a nominal exchange rate devaluation as a mechanism for short-run macroeconomic adjustment and stabilization policy in response to business cycle conditions. Clearly more theoretical and empirical work is needed to link theoretical understanding of fiscal policy in a monetary union to the actual fiscal history of these countries.

Conclusion

The experience to date of the eurozone monetary union, including the eurozone crisis, has raised new questions about the costs of giving up nominal exchange rates as a tool of adjustment. While empirical evidence paints a mixed picture of the usefulness of nominal exchange rates in promoting macroeconomic adjustment in open economies, one cannot help but wonder if macroeconomic adjustment in Greece might have been significantly easier if it had retained nominal devaluation as a policy tool.

This experience also has spurred research on creative alternatives to nominal devaluation. Fiscal devaluations are appealing in theoretical models, but there remain many questions regarding how they work and much to understand about how they can be implemented in a practical way. These theoretical developments are helpful for understanding the origins of historical experiences in the eurozone, such as the eurozone crisis, and should also be helpful in thinking about options for preventing or dealing with such circumstances in the future.