A. Trade Linkages

1. Costa Rica has substantial trade linkages, in particular, with the United States, and its exports are relatively more sophisticated than those of other CAPDR countries.² In 2013, almost 40 percent of Costa Rica exports (or 9 percent of GDP) were destined to the United States and about 30 percent (or 6 percent of GDP) were headed to other countries in the Americas—the majority to the neighboring CAPDR countries. Exports to the EU were also notable, comprising almost 20 percent of the total (4 percent of GDP), and those to Asia largely accounted for the rest. Unlike other CAPDR countries, Costa Rica’s exports, including those to the U.S. and Asia, are concentrated in knowledge-intensive products and in capital goods, attesting to its deeper integration into the global supply chains (Figure 1). Exports of food to the U.S., euro area, and other CAPDR countries are also somewhat important. The geographic distribution of imports is somewhat different. Imports from the U.S. are large—about one half of total imports (or 20 percent of GDP)—but those from Asia are also sizeable (close to 20 percent of the total or 7 percent of GDP). In contrast, imports from CAPDR, Mexico, other Latin American countries, and Europe are much less important (about 7 percent of total or 3 percent of GDP each).

View Full Size

Costa Rica: Trade Linkages

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: WITS World Bank, UNSD Comtrade, WEO, and Fund staff estimates.^1/ Other CAPDR includes Costa Rica, Honduras, Nicaragua, El Salvador, Panama and the Dominican Republic.^2/ Knowledge Intensive products include transport, electrical equipment, machinery and chemicals.^3/ PPP-weighted average.^4/ The slowdown in GDP growth in 2014 while US GDP is showing signs of recovery reflects the one-off effect of INTEL manufacturing plant withdrawal.^5/ The stages of processing include capital goods, consumer goods, intermediate goods and raw materials.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Export of Goods to U.S. as 2013

(Percent of total)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: World Bank, WITS and Fund staff estimates.

• Download Figure
• Download figure as PowerPoint slide

B. Real Growth Spillovers

2. Simple correlations reveal strong association between Costa Rica GDP growth and that of the United States as well as that of other CAPDR countries. The correlation coefficients of Costa Rica annual GDP growth with that of the U.S. as well as that of other CAPDR countries are around 0.6 each for the period 1975–2013. In fact, correlation with the U.S. is stronger for Costa Rica than for other countries in the region though it has declined over time. Nonetheless, movements in U.S. GDP growth continue to be mirrored in the movements of Costa Rica’s exports and GDP growth. Co-movements with growth in other CAPDR countries, China and Mexico have strengthened over time while those with Europe, proxied by Germany, have weakened.

3. A multi-country VAR model was used to assess the impact of growth spillovers in Costa Rica, while taking into account multilateral linkages. The model described in Poirson and Weber (2011) assesses sensitivity of Costa Rica GDP growth to growth shocks originating in its main trading partners while taking into account growth spillovers between all countries in the sample. It employs a structural VAR with the following reduced form specification

where y is obtained by stacking each country’s GDP growth rate at time t y_i,t in a vector y_t = (y_1,t…y_I,t…y_i,t) and the control vector x includes a dummy variable for global crises (oil crisis of 1979, Gulf war crisis of 1990, EMU crisis of 1991, and global financial crisis of 2008–09) as well as a constant term. The VAR was estimated on quarterly growth rates of real seasonally-adjusted GDP series for the period 1977Q1 to 2013Q4 in a sample of 6 countries/regions, namely, the United States, China, Germany, Mexico, a PPP-weighted average of 6 Central American countries (Guatemala, Nicaragua, the Dominican Republic, Panama, Honduras, and El Salvador) and Costa Rica. The VAR incorporates one lag—the optimal lag-length according to most lag-selection criteria, including the Akaike, Schwartz, and Hannan-Quinn information criteria.

4. The model was identified through the Cholesky ordering of the countries. This approach assumes that the growth rate of any country i does not depend contemporaneously on the growth rate of a country j ordered after country i. The ordering was chosen based on the results of the previous work on spillovers, including, Bayoumi and Swiston (2009) and Swiston (2010) with the U.S. ordered first and other countries’ according to their relative size, namely, China, Germany, Mexico, an average of 6 Central American countries and Costa Rica. For comparison the model was also estimated for Guatemala, El Salvador, and Honduras with the Central America averages being replaced accordingly to exclude the country of interest. We have experimented with other orderings but the results for Central America proved robust to ordering variations. The approach allows computing structural errors and coefficients that are used in the calculation of the impulse response functions. It also allows decomposing historical real GDP growth rate into the long-run, dynamic domestic and dynamic foreign components. The dynamic contributions were derived from the moving average representation of the entire history of each country’s growth rate and the structural errors, which allowed computing the contribution of the respective country’s shocks to the quarterly growth rate of GDP of the country under consideration. The long-run growth component is estimated by the constant term.

5. Two types of data interpolations were used. Since there were gaps in real GDP data for many Central American countries, we had to interpolate from annual to quarterly data to fill in those gaps. We employed two different interpolation methods to check the robustness of the results. First, we used a linear interpolation method, which involves linear projection between annual data points to substitute for the quarterly missing values. Second, we employed the method of Deeset et al. (2007), which utilizes the pattern of persistence derived from the available quarterly data to interpolate the annual series and thus generate the missing quarterly data.

6. The results suggest that shocks originating in the U.S. and the rest of Central America have a pronounced impact on Costa Rica. A 1 percentage point reduction in domestic demand growth in the U.S. over a year leads to a maximum reduction in growth in Costa Rica by 1.1 percentage points (Figure 2). Similarly, a 1 percentage point reduction in growth in other Central American countries over a year would lower Costa Rica growth by at most 0.9 percentage points. The shock to German and Chinese domestic demand growth would lower Costa Rica growth by at most ½ percentage points. These results are generally in line with the export structure as well as simple bivariate correlations. However, in contrast to simple bivariate correlations, the VAR allows tracing the shocks down to their origin. Hence, the VAR results confirm the importance of shocks originating in the U.S and CAPDR countries for Costa Rica growth. Finally, the results imply that the “external growth locomotive,” captured by the dynamic foreign component, has slowed since the 2008 financial crisis and Costa Rica may have to rely more on domestic growth sources going forward.

View Full Size

Costa Rica and Selected Trading Partners: Real Growth Spillovers

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates.

• Download Figure
• Download figure as PowerPoint slide

7. Costa Rica is generally more sensitive to external growth shocks than other countries in the region. Figure 2 demonstrates the impact on Costa Rica and Guatemala of a 1 percent over a year (four quarter shock) to the dynamic domestic growth component in 2014 in the U.S., Germany, China and other CAPDR countries. Costa Rica is generally more sensitive to external growth shocks than Guatemala, particularly to the shocks originating from the U.S. with elasticity exceeding unity. Costa Rica appears also more sensitive to external shocks than Honduras or El Salvador (the results are not shown). While other countries in the region also have substantial trade ties with the U.S., Costa Rica’s particular sensitivity to U.S. growth most likely reflects the composition of its exports. In particular, a large knowledge-intensive component of Costa Rica exports is likely to fluctuate more over the economic cycle as the demand for these goods is relatively more elastic than for basic staples that comprise a large share of the export baskets of other CAPDR countries.

8. Spillovers from Costa Rica to other CAPDR countries owing to the Intel closure are likely to be limited. The VAR model above can also be used to assess outward spillovers from Costa Rica to other countries. With the estimated impact on Costa Rica GDP growth of about ½ percentage points in 2014–15 from the closure of Intel’s manufacturing plant, growth in the neighboring CAPDR countries is estimated to decline by at most 0.2 percentage points. The impact is the largest for El Salvador and to a lesser extent for the Dominican Republic, while it is estimated to be negligible for Panama and Honduras.

View Full Size

Half Percent Point Shock in Costa Rica’s GDP Growth

(Change from baseline)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates.

• Download Figure
• Download figure as PowerPoint slide

C. Fiscal Spillovers

9. To assess the impact of fiscal spillovers from other countries to Costa Rica, a multi-country demand model was employed. The model described in Ivanova and Weber (2011) is based on the representation of the national accounts and behavioral assumptions for government spending, taxes, consumption, investment, exports and imports:

View Full Size

• Download Figure
• Download figure as PowerPoint slide

It allows simulating the impact on growth from domestic and foreign fiscal policy changes in a number of countries linked by trade. Consistent with empirical findings in the literature, fiscal measures are assumed to have an impact on GDP not only in the period, in which they are enacted, but also in the following period. The model was solved simultaneously for 20 countries, including those in CAPDR, which together account for 70 percent of the world GDP.³ The solution reflects third-country linkages among the countries in the sample as well as feedback loops between foreign and domestic fiscal policies. The approach, however, quantifies only the direct demand impact and does not reflect credibility or other non-demand driven effects to the extent that they are not embedded in the underlying multiplier estimates. Moreover, it focuses on a short-term impact (two years) and may not fully capture the effects of exchange rate and price adjustments, which are likely to reduce fiscal spillovers in the longer-term.

10. Costa Rica is more sensitive to fiscal shocks in the Americas than most CAPDR countries, with U.S. fiscal policy having the largest impact. A uniform fiscal expenditure impulse of a 1 percentage point of GDP in all countries in the sample from the Americas (the U.S., CAPDR, Mexico, Canada and Brazil) would reduce Costa Rica growth by 0.2 percentage points over two years.⁴ This is higher than the impact of a similar shock on other CAPDR countries. Each percentage point of GDP reduction in government spending in the U.S. would reduce Costa Rica growth by about 0.15 percentage points over 2 years. Fiscal spillovers from other Central and Latin America countries would have only a marginal impact.

View Full Size

Costa Rica: Fiscal Spillovers from a Uniform Fiscal Impulse in the Americas by Country

(Cumulative effect on growth over two years) ^1/

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: WEO, DOT, and Fund staff estimates.^1/ Fiscal impulse is an increase in expenditure of 1 percent of GDP in all Central American countries, the U.S., Canada, Mexico and Brazil.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Central America: Total Fiscal Spillovers from a Uniform Fiscal Impulse in the Americas

(Cumulative effect on growth over two years) ^1/

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: WEO, DOT, and Fund staff estimates.^1/ Fiscal impulse is an increase in expenditure of 1 percent of GDP in all Central American countries, the U.S., Canada, Mexico and Brazil.

• Download Figure
• Download figure as PowerPoint slide

11. Fiscal policy under the WEO baseline or that recommended by the IMF staff for 2015–16 in the Americas would have limited impact on Costa Rica growth. Under the baseline WEO scenario, the impact of fiscal policy changes in other countries on Costa Rica growth in 2015–16 is small. A small domestic fiscal relaxation in Costa Rica in 2015 is partially offset by the negative fiscal spillover from trading partner countries. With almost no expected change in domestic fiscal policy in 2016 the growth impact under the baseline is minimal. An alternative scenario incorporates IMF staff advice in terms of the total deficit reduction for 2015–16 and features an average fiscal adjustment of about 0.75 percentage points of GDP in CAPDR countries, Mexico, and Canada and an adjustment of about 2 percentage points of GDP in Costa Rica over two years. To estimate the upper bound of fiscal spillovers all the adjustment is assumed to be made through expenditure reductions, which have larger fiscal multipliers. In this alternative scenario, growth in Costa Rica would be reduced by 1 percentage points in 2015–2016, largely on account of domestic consolidation while spillovers from other countries would remain limited.

Fiscal Contribution to Growth Under the Baseline

(In percentage points)

Source: Fund staff estimates.

Fiscal Contribution to Growth Under the Baseline

(In percentage points)

		2015				2016
		Change in the Fiscal Balance	Total Growth Impact	Of which:		Change in the Fiscal Balance	Total Growth Impact	Of which:
				Domestic Effect	Spillover Effect			Domestic Effect	Spillover Effect
Costa Rica		-0.2	0.1	0.2	-0.1	0.0	0.0	0.2	-0.2
of which:
	− current year		−0.1	0.1	−0.2		0.0	0.1	−0.1
	− carry over prev. year		0.2	0.1	0.1		0.0	0.1	−0.1
PPP weighted average		0.6	−0.2	−0.2	−0.1	0.3	−0.4	−0.3	−0.1

Source: Fund staff estimates.

View Table

Fiscal Contribution to Growth Under the Baseline

(In percentage points)

		2015				2016
		Change in the Fiscal Balance	Total Growth Impact	Of which:		Change in the Fiscal Balance	Total Growth Impact	Of which:
				Domestic Effect	Spillover Effect			Domestic Effect	Spillover Effect
Costa Rica		-0.2	0.1	0.2	-0.1	0.0	0.0	0.2	-0.2
of which:
	− current year		−0.1	0.1	−0.2		0.0	0.1	−0.1
	− carry over prev. year		0.2	0.1	0.1		0.0	0.1	−0.1
PPP weighted average		0.6	−0.2	−0.2	−0.1	0.3	−0.4	−0.3	−0.1

Source: Fund staff estimates.

Fiscal Contribution to Growth Under the Recommended Fiscal Path

(In percentage points)

Source: Fund staff estimates.

Fiscal Contribution to Growth Under the Recommended Fiscal Path

(In percentage points)

		2015				2016
		Change in the Fiscal Balance	Total Growth Impact	Of which:		Change in the Fiscal Balance	Total Growth Impact	Of which:
				Domestic Effect	Spillover Effect			Domestic Effect	Spillover Effect
										Costa Rica		1.0	-0.4	-0.2	-0.1	1.0	-0.7	-0.5	-0.2
of which:
	− current year		−0.5	−0.4	−0.2		−0.4	−0.3	−0.1
	− carry over prev. year		0.2	0.1	0.1		−0.3	−0.2	−0.1
PPP weighted average		0.6	−0.2	−0.2	−0.1	0.4	−0.4	−0.3	−0.1

Source: Fund staff estimates.

View Table

Fiscal Contribution to Growth Under the Recommended Fiscal Path

(In percentage points)

		2015				2016
		Change in the Fiscal Balance	Total Growth Impact	Of which:		Change in the Fiscal Balance	Total Growth Impact	Of which:
				Domestic Effect	Spillover Effect			Domestic Effect	Spillover Effect
										Costa Rica		1.0	-0.4	-0.2	-0.1	1.0	-0.7	-0.5	-0.2
of which:
	− current year		−0.5	−0.4	−0.2		−0.4	−0.3	−0.1
	− carry over prev. year		0.2	0.1	0.1		−0.3	−0.2	−0.1
PPP weighted average		0.6	−0.2	−0.2	−0.1	0.4	−0.4	−0.3	−0.1

Source: Fund staff estimates.

D. Financial Spillovers

12. To assess the impact of financial spillovers to Costa Rica from stress in international banks, the IMF Bank Contagion Module was used.⁵ This module, based on BIS banking statistics and bank-level data, estimates potential rollover risks for Costa Rica stemming from both foreign banks’ affiliates operating in Costa Rica and foreign banks’ direct cross-border lending to Costa Rica borrowers. Rollover risks were triggered in the scenarios analyzed here by assuming bank losses in the value of private and public sector assets in certain countries and/or regions. If the banks do not have sufficient capital buffers to cover the losses triggered in a given scenario, they have to deleverage (reduce their foreign and domestic assets) to restore their capital-to-asset ratios,⁶ thus squeezing credit lines to Costa Rica and other countries. The estimated impact on losses in cross-border credit availability for Costa Rica also incorporates the transmission of shocks through Panama, given its central financial role in the region. The assumption is that cross-border lending from Panama to Costa Rica declines proportionally to the decline in cross-border lending to Panama from the banking systems where the shocks originate.⁷

13. Spillovers to Costa Rica from stress in international banks are moderate. The impact on foreign credit availability in Costa Rica of the severe stress scenarios in asset values of BIS reporting banks, presented in the table below and in the text figure, is larger than in other countries in the region with the exception of Panama and El Salvador. For example, spillovers from a 10 percent shock to assets originating in the U.S. and Canada would reduce credit in Costa Rica by almost 4 percent of GDP (or 6 percent of total domestic and cross-border credit to the public and private sectors). In contrast, a similar shock would reduce credit in Guatemala by only 0.6 percent of GDP (or 1 percent of total). More generally, the level of upstream exposures of Costa Rica to international banks⁸ is moderate with the upper limit of rollover risks on external credit of about 13 percent of GDP (or 20 percent of total domestic and cross-border credit to the public and private sectors in Costa Rica).⁹ This upper limit would correspond to a worst case scenario without any replacement, either domestic or external, of the loss of credit by BIS reporting banks to Costa Rican borrowers.

View Full Size

Spillovers from International Banks’ Exposures as of September 2013: Effect on Foreign Credit of 10% Loss in All Bank Assets of BIS-Reporting Banks

(Percent of GDP)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: BIS and Fund staff estimates.

• Download Figure
• Download figure as PowerPoint slide

14. Spillovers from a shock originating in the U.S. assets only are even more moderate, but financial regional integration is important in the transmission of shocks. Despite the large share of U.S. banks in total foreign bank claims on Costa Rica, the impact of U.S. asset losses in U.S. and international banks on cross-border credit availability in Costa Rica—0.3 percent of GDP in response to a 10 percent loss in U.S. asset values—is limited. This effect has declined in recent years reflecting both the strengthening in international banks’ capital buffers and the cross-border deleveraging of assets after the global financial crisis. As of September 2013, the most sizable impact on claims on Costa Rican borrowers would stem from shocks in Europe. A 10 percent loss on European assets would result in a reduction in credit availability to Costa Rica of about 4 percent of GDP. This result is also driven by the increasing importance of financial integration with other countries in the region. Indeed, almost one third of the estimated credit losses in Costa Rica resulting from a shock originating in Europe would be transmitted through cross-border lending from Panama, which is more dependent on European banks’ funding.

Spillovers to Costa Rica from International Banks’ Exposures as of September 2013 ^1/

Source: RES/MFU Bank Contagion Module based on BIS, ECB, and IFS data.

^1/In addition to BIS-reporting banks, Fund staff analyzed spillovers through Panama.

^2/Magnitude denotes the percent of on-balance sheet claims (all borrowing sectors) that default.

^3/Reduction in foreign banks credit to Guatemala due to the impact of the analyzed shock on their balance sheets, assuming a uniform deleveraging across domestic and external claims.

^4/Greece, Ireland, and Portugal.

^5/Greece, Ireland, Portugal, Italy, Spain, France, Germany, Netherlands, and UK.

Spillovers to Costa Rica from International Banks’ Exposures as of September 2013 ^1/

	Shock Originating From Magnitude 2/	Impact on claims on CRI borrowers (percent of GDP) 3/
USA	10	−0.29
Canada	10	−3.48
USA and Canada	10	−3.92
UK	10	−1.77
Switzerland	10	−0.13
Germany	10	−1.35
France	10	−0.33
Spain	10	−0.09
Spain	30	−0.09
Netherlands	10	−0.33
Italy	10	−0.04
Italy	30	−0.05
Greece	10	−0.07
Portugal	10	−0.02
GIP 4/	10	−0.09
GIP 4/	30	−0.09
Japan	10	−0.74
Korea	10	0.00
Taiwan	10	0.00
European Countries 5/	10	−4.04

Source: RES/MFU Bank Contagion Module based on BIS, ECB, and IFS data.

^1/In addition to BIS-reporting banks, Fund staff analyzed spillovers through Panama.

^2/Magnitude denotes the percent of on-balance sheet claims (all borrowing sectors) that default.

^3/Reduction in foreign banks credit to Guatemala due to the impact of the analyzed shock on their balance sheets, assuming a uniform deleveraging across domestic and external claims.

^4/Greece, Ireland, and Portugal.

^5/Greece, Ireland, Portugal, Italy, Spain, France, Germany, Netherlands, and UK.

View Table

Spillovers to Costa Rica from International Banks’ Exposures as of September 2013 ^1/

	Shock Originating From Magnitude 2/	Impact on claims on CRI borrowers (percent of GDP) 3/
USA	10	−0.29
Canada	10	−3.48
USA and Canada	10	−3.92
UK	10	−1.77
Switzerland	10	−0.13
Germany	10	−1.35
France	10	−0.33
Spain	10	−0.09
Spain	30	−0.09
Netherlands	10	−0.33
Italy	10	−0.04
Italy	30	−0.05
Greece	10	−0.07
Portugal	10	−0.02
GIP 4/	10	−0.09
GIP 4/	30	−0.09
Japan	10	−0.74
Korea	10	0.00
Taiwan	10	0.00
European Countries 5/	10	−4.04

Source: RES/MFU Bank Contagion Module based on BIS, ECB, and IFS data.

^1/In addition to BIS-reporting banks, Fund staff analyzed spillovers through Panama.

^2/Magnitude denotes the percent of on-balance sheet claims (all borrowing sectors) that default.

^3/Reduction in foreign banks credit to Guatemala due to the impact of the analyzed shock on their balance sheets, assuming a uniform deleveraging across domestic and external claims.

^4/Greece, Ireland, and Portugal.

^5/Greece, Ireland, Portugal, Italy, Spain, France, Germany, Netherlands, and UK.

E. Monetary Policy Spillovers (the Effect of U.S. Monetary Policy Normalization)

15. The exit from U.S. unconventional monetary policy presents new challenges for Costa Rica. While faster-than-expected growth in the U.S. would help stimulate exports, the withdrawal of the monetary stimulus by the Federal Reserve could lead to a faster tightening of global financial conditions and, possibly, higher market volatility as global risk appetite declines. Tighter financial conditions would imply higher financing costs for the government and the private sector with negative implications for domestic investment and consumption. These developments may be particularly relevant for Costa Rica due to its substantial economic ties with the U.S.

16. The note employs two approaches to analyzing the impact of U.S. tapering on the economy of Costa Rica. First, we use an empirical approach to assess the effect of changes in U.S. monetary policy on external government financing costs and GDP growth in Costa Rica. Second, we employ a system of general equilibrium models to simulate the impact of faster growth in the U.S., accompanied by tighter global financial conditions, on Costa Rica’s GDP, inflation, current account and other macroeconomic variables under various scenarios.

1. Empirical Approach

EMBI Spreads Model

17. A regression-based approach was used to assess the effects of U.S. tapering on external government financing costs. The model was inspired by Evan Papageorgiou’s work in Chapter 1 of the April 2013 Global Financial Stability Report. Specifically, a random effects panel regression model was estimated linking the spread between the yield on foreign currency government bonds and U.S. bonds (EMBI spread) to domestic and external conditions at monthly frequency in a sample of Latin American countries,¹⁰ namely,

${EMBI}_{it}={\alpha }^{*}IN{T}_t^{US}+{\beta }^{*}{VULN}_i+{\gamma }^{*}IN{T}_t^{US}*{VULN}_i+{\eta }^{*}{ICRG}_{it}+{\delta }_i+{\varepsilon }_{it}$

where EMBI_it s EMBI spread of country i at time t, $IN{T}_t^{US}$ is the level of U.S. real interest rate (the yield on U.S. Treasury Inflation-Protected Securities, TIPS), which captures global financial conditions; VULN_i is a time-invariant index of domestic vulnerabilities constructed by the team, which captures domestic economic fundamentals that are not changing substantially from month-to-month (see below); ICRG_it is a time-varying index of domestic political, financial and economic risk from International Country Risk Guide (ICRG),¹¹ which captures short-term movements in domestic conditions. δ_i is a country’s random effect, which is assumed to be uncorrelated with the other regressors. We have also analyzed the impact of U.S. “money” and “real” shocks separately using a structural VAR approach with sign restrictions described in IMF (2014), which includes the long-term bond yields and log of stock prices.

18. A measure of domestic vulnerabilities is an important explanatory variable in the empirical model. It is captured by a time-invariant index, the values of which are summarized below in the text table. The index is based on 21 economic indicators that were found important is predicting the risk of capital flow reversal as well as movements in asset prices.¹² The indicators were grouped in 7 major categories, including a measure of the size of the recent capital inflows, external balance and debt, the amount of local debt securities held by foreigners, reserve cushion, fiscal position as well as banking system and private sector leverage. Each indicator was compared to two thresholds, either found important in earlier studies or based on the standard deviations in the sample, to assign a score of 0, 1 or 2 with 0 capturing les vulnerability. Then the index was constructed based on the weighted average of these scores, with 55 percent weight put on indicators of external vulnerability, which have been found particularly important in the literature, 20 percent weight on the fiscal position, 15 percent weight on banking vulnerabilities, and 10 percent weight on the share of local debt securities held by foreigners.

	Weight in total indicator	Central America							LA5 (PPP-weighted average)
		CRI	DR	GTM	HND	NIC	PAN	SLV
Size of recent capital inflows	0.10	1	1	0	1	2	2	0	0.8
External balance	0.15	1	1	1	1	2	1	1	1.1
External Debt	0.10	1	0	0	0	1	2	1	0.4
Debt securities held by foreigners	0.10	0	0	0	0	0	0	0	0.8
Gross reserve cushion	0.20	0	2	0	2	0	1	1	0.0
Fiscal position	0.20	2	1	0	2	0	0	2	0.8
Banking system and private sector leverage	0.15	1	0	1	1	0	2	1	0.6
Overall	1.00	0.9	0.9	0.3	1.2	0.6	1.1	1.0	0.6

View Table

	Weight in total indicator	Central America							LA5 (PPP-weighted average)
		CRI	DR	GTM	HND	NIC	PAN	SLV
Size of recent capital inflows	0.10	1	1	0	1	2	2	0	0.8
External balance	0.15	1	1	1	1	2	1	1	1.1
External Debt	0.10	1	0	0	0	1	2	1	0.4
Debt securities held by foreigners	0.10	0	0	0	0	0	0	0	0.8
Gross reserve cushion	0.20	0	2	0	2	0	1	1	0.0
Fiscal position	0.20	2	1	0	2	0	0	2	0.8
Banking system and private sector leverage	0.15	1	0	1	1	0	2	1	0.6
Overall	1.00	0.9	0.9	0.3	1.2	0.6	1.1	1.0	0.6

19. Costa Rica appears relatively more vulnerable compared to other Latin American countries due to both domestic and external vulnerabilities. This reflects largely the high fiscal deficit and rapidly rising public debt. However, elevated portfolio inflows in 2012, elevated short-term external debt as well as a high current account deficit add somewhat to the scope of vulnerabilities, though the high share of FDI financing moderates external risks. The recent expansion of private credit with an increased reliance on bank foreign funding is an additional source of vulnerability.

20. Empirical model estimates suggest that the response of Latin America bond spreads to the tightening of global financial conditions depends on the degree of domestic vulnerabilities. The increase in the EMBI spread in response to an increase in the U.S. real interest rate is more pronounced for countries that are more vulnerable to capital flow reversal. In particular, an increase in U.S. real interest rates by 100 basis points is associated with an increase in EMBI spread of about 20 basis points in a country that exhibits low vulnerability (vulnerability index=0). A similar increase in the U.S. real rates in a highly vulnerable country (vulnerability index=2) would be associated with an increase in EMBI spread of about 75 basis points. In addition, domestic vulnerabilities directly contribute to the level of the EMBI spread, with highly vulnerable countries having a spread of about 430 basis points higher than countries with low vulnerability. Moreover, a reduction in domestic risks, captured by the ICRG index, might help reduce the EMBI spread, though this result was not robust across the different econometric specifications. The estimates from the model separating U.S. interest rate shock into “money” and “real” shocks also suggest a differential impact. “Money” shocks work to increase the EMBI spread, while “real” shocks work to decrease the spread.

Dependent Variable= EMBI spread	Random effects	Fixed effects
ICRG	−10.3340***	−4.2176
	[-4.01]	[-1.47]
INTUS	0.1824**	0.2146***
	[2.23]	[2.83]
INTUS*VULN	0.2882***	0.2753***
	[2.99]	[3.09]
VULN	213.8781***
	[5.88]
Observations	140	140
Between R-squared	0.75	0.68
Number of countries	11	11

View Table

Dependent Variable= EMBI spread	Random effects	Fixed effects
ICRG	−10.3340***	−4.2176
	[-4.01]	[-1.47]
INTUS	0.1824**	0.2146***
	[2.23]	[2.83]
INTUS*VULN	0.2882***	0.2753***
	[2.99]	[3.09]
VULN	213.8781***
	[5.88]
Observations	140	140
Between R-squared	0.75	0.68
Number of countries	11	11

21. Thus, Costa Rica’s relatively weak macroeconomic fundamentals, in particular, on the fiscal side, could amplify the transmission of the global shocks. With the index of domestic vulnerability for Costa Rica estimated at 0.9 on a scale from 0 to 2, the increase in real U.S. interest rates by about 100 basis points that occurred between May 2013 and January 2014, is estimated to have contributed to a widening of Costa Rica EMBI spread by about 50 basis points. Overall, according to the model, Costa Rica’s EMBI spread should have increased by about 60 basis points, taking into account a small deterioration in the domestic vulnerability component since May 2013. In fact, Costa Rica’s EMBI spread widened even more during this period (by about 100 basis points) and the increase was larger than in other countries in the region, possibly reflecting a higher weight that investors put on the weaknesses in fiscal situation. In fact, Costa Rica’s EMBI spread remained elevated in H1 of 2014 when the increases in the spreads of other Latin America countries largely reverted to pre-tapering levels.

View Full Size

Central America: Contribution to Changes in EMBI spread

(May 2013 - January 2014)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates

• Download Figure
• Download figure as PowerPoint slide

2. VAR Growth Model

22. Empirical estimates suggest that accelerated monetary policy tapering in the U.S. is likely to have a positive impact on GDP growth in Costa Rica. A VAR approach was used to analyze the impact of U.S. growth and real interest rate shocks on Costa Rica GDP growth. A VAR model described in the section B above was augmented with the U.S. real 10 year government bond yield as an endogenous variable. The results suggest that an increase in U.S. real demand growth by 1 percentage point over 4 quarters would raise Costa Rica GDP growth by at most 1.2 percentage points. An increase in U.S. real bond yield by 1 percentage points over 4 quarters would reduce Costa Rica GDP growth by 0.4 percentage points. Hence, on balance the impact is likely to be positive.

General Equilibrium Model Simulations

23. A general equilibrium model was used to assess the impact of U.S. tapering on Costa Rica’s GDP and other macroeconomic variables. The full impact of the U.S. tapering on the Costa Rican economy was assessed using a module of the IMF’s Flexible System of Global Models (FSGM). It comprises a system of multi-region, general equilibrium models combining micro-founded and reduced-form relationships for various economic sectors. The model has a fully articulated demand side, and some supply side features. International linkages are modeled in aggregate for each country/region. The level of public debt in each country and the resulting implications for national savings determine the global real interest rate in the long run. The parameters of the model, except those determining the cost of adjustment in investment, have been estimated from the data using a range of empirical techniques. Real GDP in the model is determined by the sum of the components of demand in the short run and the level of potential output in the long run. The households’ consumption-savings decisions are explicitly micro founded as are firms’ investment decisions. Government absorption is determined exogenously, while imports and exports are specified with reduced-form models.

View Full Size

Costa Rica: Impact on Real GDP Growth from U.S. Real Demand Growth and Real Interest Rate Shocks ^1/

(Percent points deviation from baseline)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates.^1/ U.S. real interest rate used in VAR estimation is U.S. real 10 year government bond yield.

• Download Figure
• Download figure as PowerPoint slide

24. The model incorporates endogenous rules governing the operation of fiscal and monetary policy. In particular, the government sector, besides government absorption, encompasses additional spending by the fiscal authority in the form of lump-sum transfers to all households, or targeted exclusively to liquidity constrained households. The fiscal authority chooses a long-run level of debt relative to GDP. In order to meet its debt and deficit targets as well as spending obligations, the governments have recourse to the following tax instruments: consumption taxes, labor income taxes, corporate income taxes and lump-sum taxes. In the presence of shocks to the economy, all tax rates remain fixed and spending on general lump-sum transfers adjusts to ensure that the public debt-to-GDP ratio is maintained in the medium term. Monetary policy is governed by an endogenous rule, which features interest rate smoothing and responds to the deviation of inflation from its target, the deviation of expected inflation from its target, as well as the output gap. Inflation is modeled via a reduced form Phillips’ curve with inflation being a function of expected inflation (lag and model-consistent lead), as well as the output gap, and the change in the log of the real effective exchange rate. The exchange rate in the short run is determined via the uncovered interest parity, while in the long run it adjusts to ensure external stability given households desired holdings of net foreign assets.

25. The simulations included 4 scenarios. Scenario 1 (S1) features a U.S. growth surprise of about 1 percent by 2015 relative to the baseline (Figures 1 and 2). Scenario 2 (S2) includes faster U.S. growth as in Scenario 1 with U.S. interest rates rising by about 80 basis points, but this is accompanied by an increase in the risk premium outside the U.S., with market interest rates rising by about 1 standard deviation in each Latin America country. The risk-premium shock for Costa Rica amounts to an increase of about 100 basis points. Scenario 3 (S3) incorporates, in addition to the assumptions in Scenario 2, an increase in the U.S. risk premium, with the rates rising by about 40 basis points at the peak. Finally, Scenario 4 (S4) incorporates, in addition to the assumptions in Scenario 2, a fiscal consolidation of about ½ percentage points of GDP in the next six years through an increase in consumption taxes. Fiscal consolidation is assumed to essentially undo the increase in risk premium in Scenario 2.

26. The results indicate that the impact of U.S. tapering on Costa Rica GDP is likely to be positive in the short run. In Scenario 1, a U.S. growth surprise of about 1 percent by 2015 relative to the baseline (Figure 1, S1) results in higher GDP in Costa Rica by about 0.6 percent compared to the baseline (Figure 2, S1). Faster growth in the U.S. stimulates Costa Rica exports while domestic consumption remains largely unchanged. The decline in investment due to higher domestic interest rates, raised in response to output exceeding its potential, is also relatively small. In contrast, in Scenario 2 the acceleration of U.S. growth is accompanied by an increase in the risk premium, with a substantial increase in market interest rates for Costa Rica. In this scenario, Costa Rican GDP remains essentially unaffected—an increase of only 0.1 percent compared to the baseline—as stronger demand for Costa Rica exports from the U.S. is offset by weaker domestic investment and consumption due to tighter financial conditions (Figures 1 and 2, S2). Hence, even this rather severe scenario implies a relatively benign outcome for Costa Rica. Finally, in Scenario 3 with the perception of a more “hawkish” U.S. Fed, which is modeled through an increase in the U.S. risk premium and, consequently, a somewhat weaker U.S. growth, the positive impact on Costa Rican GDP is reduced further, albeit only slightly (Figure 2, S3). All 3 scenarios suggest that the longer-term effect could be slightly negative as lower investment due to tighter financing conditions will lead to a temporarily lower potential GDP in Costa Rica and the cyclical rise in U.S. imports dissipates as prices and exchange rates adjust and U.S. output returns to its potential.

View Full Size

The Impact of U.S. Tapering on Costa Rica’s GDP

(Deviation from baseline)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Sources: Fund staff estimates.

• Download Figure
• Download figure as PowerPoint slide

27. In addition, the results suggest that Costa Rica’s fiscal sustainability objectives can be safely achieved in the context of U.S. tapering. Specifically, in Scenario 4, with faster growth in the U.S. and accelerated monetary policy normalization accompanied by an increase in risk premium outside the U.S., while Costa Rica pursues a fiscal tightening of about 1 percentage points of GDP in the next three years, the country can lower the public debt-to-GDP ratio by close to 20 percentage points by 2018 compared to the baseline where the debt is expected to rise by about the same amount, without a significant negative impact on growth. In fact, growth accelerates as the risk premium is reduced substantially, thereby supporting investment, while savings from lower interest payments allow the government to raise transfers, thus supporting private consumption.

View Full Size

Costa Rica: GDP

(Percent deviation from baseline)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Costa Rica: Debt

(Percentage points of GDP deviation from baseline)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates

• Download Figure
• Download figure as PowerPoint slide

Table 1.

Costa Rica: Vulnerability Indicators

Table 1.

Costa Rica: Vulnerability Indicators

Category	Indicator	Period	Units	Thresholds	Less vulnerable if the indicator is
Size of recent capital inflows	Cumulative capital inflows proxied by the balance on the financial account	2010-12	% of 2012 GDP	14; 26	Lower
	Cumulative portfolio and other investment flows	2010-12	% of 2012 GDP	4:11	Lower
External balance	Current account balance	2012	% of GDP	−3; +2	Higher
	Change in the current account balance	2007-12	% of GDP	−3;0	Higher
	Current account balance plus FDI	2012	% of GDP	−3; +2	Higher
External debt	External Debt	2012	% of GDP	30; 70	Lower
	Change in external debt	2007-12	% of GDP	1; 9	Lower
	Short-term external debt	2012	% of GDP	5; 20	Lower
	Change in short-term external debt	2007-12	% of GDP	0; 5	Lower
Local debt securities held by non-residents	Proportion of domestic government securities held by foreigners	2012	%	20; 30	Lower
Gross reserve cushion	Reserve as percent of metric	2013	%	75; 100	Higher
Fiscal position	Primary balance	2012	% of GDP	−3; -1	Lower
	Government debt	2012	% of GDP	35; 60	Lower
	Change in government debt	2007-12	% of GDP	0; 5	Lower
	Expected change in government debt	2012-18	% of GDP	0; 5	Lower
Banking system and private sector leverage	Credit to the private sector	2012	% of GDP	40; 65	Lower
	Change in credit to the private sector	2007-12	% of GDP	3; 18	Lower
	Share of credit expansion financed by increase in bank liabilities to non-residents	January 2011	%	25; 50	Lower
	Correlation of changes in private credit and foreign liabilities	to April 2013 2010M1-2013M6	Units	0.7; 0.9	Lower
	Loan-to-deposit ratio	2012	%	80; 100	Lower
	Change in loan-to-deposit ratio	2007-12	%	0; 19	Lower

View Table

Table 1.

Costa Rica: Vulnerability Indicators

Category	Indicator	Period	Units	Thresholds	Less vulnerable if the indicator is
Size of recent capital inflows	Cumulative capital inflows proxied by the balance on the financial account	2010-12	% of 2012 GDP	14; 26	Lower
	Cumulative portfolio and other investment flows	2010-12	% of 2012 GDP	4:11	Lower
External balance	Current account balance	2012	% of GDP	−3; +2	Higher
	Change in the current account balance	2007-12	% of GDP	−3;0	Higher
	Current account balance plus FDI	2012	% of GDP	−3; +2	Higher
External debt	External Debt	2012	% of GDP	30; 70	Lower
	Change in external debt	2007-12	% of GDP	1; 9	Lower
	Short-term external debt	2012	% of GDP	5; 20	Lower
	Change in short-term external debt	2007-12	% of GDP	0; 5	Lower
Local debt securities held by non-residents	Proportion of domestic government securities held by foreigners	2012	%	20; 30	Lower
Gross reserve cushion	Reserve as percent of metric	2013	%	75; 100	Higher
Fiscal position	Primary balance	2012	% of GDP	−3; -1	Lower
	Government debt	2012	% of GDP	35; 60	Lower
	Change in government debt	2007-12	% of GDP	0; 5	Lower
	Expected change in government debt	2012-18	% of GDP	0; 5	Lower
Banking system and private sector leverage	Credit to the private sector	2012	% of GDP	40; 65	Lower
	Change in credit to the private sector	2007-12	% of GDP	3; 18	Lower
	Share of credit expansion financed by increase in bank liabilities to non-residents	January 2011	%	25; 50	Lower
	Correlation of changes in private credit and foreign liabilities	to April 2013 2010M1-2013M6	Units	0.7; 0.9	Lower
	Loan-to-deposit ratio	2012	%	80; 100	Lower
	Change in loan-to-deposit ratio	2007-12	%	0; 19	Lower

View Full Size

View Full Size

United States: Three Scenarios of Tapering Off (FSGM simulations)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates.

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

View Full Size

View Full Size

View Full Size

Costa Rica: The Impact of U.S. Tapering Off (FSGM simulations)

Citation: IMF Staff Country Reports 2015, 030; 10.5089/9781475527025.002.A002

Source: Fund staff estimates.

• Download Figure
• Download Figure
• Download Figure
• Download figure as PowerPoint slide