Selected Issues

Abstract

Selected Issues

Panama: Growth at risk1

Accommodative financial conditions support economic growth in the near-term, but can contribute to the build-up of financial imbalances overtime and potentially put economic growth at risk. This paper assesses financial conditions in Panama using the growth at risk approach (GaR) to link financial conditions to the distribution of future growth outcomes for Panama.2 While financial conditions in Panama remain broadly accommodative and should continue to support near-term growth prospects, the GaR model shows that the prolonged period of accommodative financial conditions can contribute to the build-up of financial vulnerabilities, putting at risk financial stability and growth over the medium-term. These results highlight the importance of remaining vigilant and building resilience to emerging financial risks. In the context of Panama’s dollarized economy with no central bank, macroprudential policy and crisis preparedness/management have an enhanced role to play in both mitigating and managing these risks.

A. Introduction

1. Financial conditions and economic activity are closely intertwined. While accommodative financial conditions support near-term growth prospects, financial vulnerabilities tend to accumulate over prolonged periods of financial excess, entailing significant downside risks for the economy over the medium-term. This paper examines the empirical relationship between financial conditions and economic activity in Panama using the growth at risk (GaR) approach. The GaR approach considers how changes in financial conditions signal risks to future GDP growth at different time horizons. The paper first provides an overview of financial conditions in Panama along three dimensions: (i) the price of risk, (ii) leverage, and (iii) external conditions. Financial conditions are then mapped into the probability distribution of future GDP growth at different forecast horizons to evaluate how different dimensions of financial conditions affect risks to the near- and medium-term growth outlook for Panama.

B. Financial Conditions in Panama

2. Financial conditions indices (FCI) are estimated to capture recent movements in financial conditions in Panama. An aggregate FCI is estimated together with separate FCIs for three important dimensions of financial risk: (i) the price of risk, (ii) leverage, and (iii) external conditions. All FCIs are estimated using principal component analysis (PCA), an approach that aggregates information about the common trend among financial indicators. In total, a set of 25 financial indicators is considered in the aggregate FCI. These indicators are then partitioned into groups to estimate the FCIs for the subdimensions of financial conditions. Using this partitioning approach, movements in the price of risk are captured by changes in interest rates, asset returns and price volatility, while movements in leverage are captured by those in credit aggregates and growth. External conditions are themselves separated into two subdimensions: (i) financial conditions, reflecting global risk sentiment and interest rates; and (ii) external demand, reflecting movements in growth in key trading partners, world trade, and commodity prices (Table 1). The estimated FCIs are normalized around zero, such that higher positive FCIs indicate relatively tighter financial conditions, and higher negative FCIs indicate more accommodative financial conditions.3

Table 1.

Panama: Financial Variables by Dimension of Financial Risk

article image
Source Author’s calculations.

3. Aggregate financial conditions in Panama remain accommodative. Movements in the aggregate FCI are broadly intuitive, suggesting relatively tighter financial conditions during the global financial crisis, followed by a period of sustained accommodative financial conditions. The advantage of the aggregate FCI is that it combines the information from the various dimensions of financial conditions to have an overall view on such conditions. However, by construction, the aggregate index may be dominated by the price of risk variables given the larger number of these variables included relative to variables capturing the other dimensions of financial conditions because of data constraints, particularly with respect to leverage variables, and suppress valuable information from the other dimensions of financial conditions for risks to the growth outlook. Indeed, the weights or loadings of the financial variables included in the aggregate FCI are dominated by the price of risk variables as well external conditions, while leverage variables receive relatively small weights. As the price of risk, leverage, and external financial conditions can affect the distribution of the growth outlook differently at different horizons (see IMF 2017), the aggregate FCI for Panama may mask the importance of the leverage or credit cycle for risks to growth in Panama, particularly at different forecast horizons. Therefore, distinct FCIs are also estimated for the main dimensions of financial risk.

uA03fig01

Panama: Aggregate FCI

(Index)

Citation: IMF Staff Country Reports 2019, 012; 10.5089/9781484394069.002.A003

Sources: Author calculations. Higher values indicate tighter financial conditions.
uA03fig02

Panama: Aggregate FCI Variable Loadings

Citation: IMF Staff Country Reports 2019, 012; 10.5089/9781484394069.002.A003

Sources: Author calculations.

4. The different FCIs suggest that the subdimensions of financial risk do not always move together:

  • The FCI related to the price of risk indicates that conditions have remained broadly accommodative since early 2010. While banks’ funding costs have started to rise with global interest rates, notably the ongoing gradual normalization of U.S. monetary policy, upward pressure on the price of risk in Panama has thus far been contained. While lending rates are primarily variable and reset automatically with movements in LIBOR, the overall increase in the price of risk has been offset by historically low sovereign and corporate spreads and continued growth in equity prices. Overall, the FCI for the price of risk captures the tightening during the global financial crisis and subsequent loosening and largely mirrors developments in the aggregate FCI, confirming that movements in the aggregate FCI are dominated by the price of risk.

  • The FCI related to leverage cycle has started to turn. While the estimated FCI capturing leverage was largely accommodative during the post-global financial crisis credit boom, it has tightened with the slowdown in credit growth since 2015. Nevertheless, on balance, the FCI capturing leverage suggests that leverage is now broadly neutral. The estimation of the leverage FCI is; however, hindered by a lack of detailed information on leverage in Panama. With financial intermediation largely bank-based and relatively shallow capital markets, sufficient time-series data was not available to include variables like equity or bond market capitalization to capture developments in leverage stemming from capital markets. Similarly, detailed data on credit from the banking system is available on a quarterly basis beginning only in 2003, hindering the inclusion of lengthy time series of credit aggregates or credit growth. Finally, quarterly national accounts data is available only beginning in 2007, limiting the period over which the credit-to-GDP ratio can be included.4 As a result, the approach includes only the credit-to-GDP ratio and growth in credit to the private sector, with the weights assigned to both variables equivalent.

  • Panama’s FCI related to external conditions indicates that these are broadly neutral. The FCI capturing external financial conditions accurately captures the sharp tightening of financial conditions during the global financial crisis and subsequent period of accommodation, while that for external demand accurately captures related movements in the global business cycle. While the FCIs suggest that, on balance, external conditions remain broadly neutral, that for external financial conditions captures the recent tightening consistent with the normalization of global interest rates.

Figure 1.
Figure 1.
Figure 1.

Panama: FCIs for the Subdimensions of Financial Risk and Variable Loadings

Citation: IMF Staff Country Reports 2019, 012; 10.5089/9781484394069.002.A003

C. Growth and Financial Conditions

5. Financial conditions contain important information about the probability distribution of future growth outcomes. The conditional density forecast of future GDP growth in Panama based on current financial and external conditions is estimated using quantile regressions following the approach in IMF (2017). 5 The quantile regressions regress future GDP growth (yt+h), on current growth (yt), financial conditions, and external demand:

Q(yt+h,q)=βyqyt+βpqpt+βlqlt+βffqext_fint+βfdqext_demt+εt,h

where q indicates the quantile level and h the forecast horizon (in quarters). The regression is fitted on a set of quantiles (0.10, 0.25, 0.50, 0.75, 0.90) and for forecast horizons of 4, 8 and 12 quarters to consider the impact of financial conditions on growth density forecasts at different horizons. The price of risk (pt), leverage (lt), external financial conditions (ext_fint), and external demand (ext_demt) as estimated through the corresponding FCIs, are included separately in the quantile regressions to investigate the relative significance of each dimension of financial conditions for signaling risks to the near- and medium-term growth outlook.

6. Results of the quantile regressions suggest that the different dimensions of financial conditions have divergent impacts on the growth forecast depending on the forecast horizon. External financial conditions are the main driver of Panama’s short-term growth prospects, while the build-up of financial vulnerabilities related to leverage is the most important link between financial conditions and Panama’s medium-term growth outlook:

  • The impact of the price of risk on the growth outlook is difficult to disentangle for Panama. The results of the quantile regressions suggest that a rising price of risk is (surprisingly) consistent with upside risks to the growth outlook over both the near- and medium-term, while a rising price of risk has typically been found to be associated with downside risks to growth, particularly over the short-term.6 However, the period for which the quantile regressions are estimated provides important insights into this result. Data constraints restrict the estimation period to start only in 2004 and, for much of this period, the price of risk has been positively, rather than inversely, correlated with real GDP growth. This positive correlation may be due to a decoupling of the business cycles of Panama and the rest of the world to the extent to which the domestic price of risk has been driven by external financial conditions. It may also be related to the completion of several substantial infrastructure projects, including the completion of the Panama Canal, that boosted growth significantly and wound down over the same period as the price of risk was becoming more accommodative.7 Therefore, the finding that a higher price of risk has a large and positive impact on the right quantiles of GDP growth could be a consequence of higher demand for capital over this investment boom in the upswing that pushed up the cost of capital. On balance, the results should not be inferred to suggest that a rising price of risks is not a useful signal of downside risks to the growth outlook, particularly as, prior to the recent period of significant investment in large-scale infrastructure projects, a rising price of risk was historically negatively correlated with growth outcomes.

  • Leverage has a small effect on growth at short-term horizons, but a negative effect at longer horizons that dominates the effect of the price of risk and external conditions. Over the short-term, leverage in Panama has a relatively neutral/slight positive effect on the growth outlook, consistent with the demand side effect of leverage dominating in the short-term with higher leverage translating into more economic activity (e.g. IMF (2017)). However, this result may be skewed by the data limitations outlined above. Over longer horizons; however, higher leverage negatively affects growth, particularly the left-hand tail of the growth distribution, as higher leverage leads to the build-up of balance sheet vulnerabilities over time. Macroprudential policy can therefore play an important role in mitigating medium-term risks to the growth outlook from the excessive build-up of leverage, with development of a framework and tools for macroprudential policy particularly imperative to provide sufficient policy flexibility to address related macro-financial risks in the context of Panama’s dollarized economy and regional financial center. This finding highlights the term structure of GaR from financial conditions: in the near-term Panama’s still broadly accommodative financial conditions will likely continue to support economic activity, but over the medium-term the continued build-up of financial vulnerabilities can shift the distribution of future GDP growth, increasing GaR.

  • External conditions are the main drive of Panama’s growth over short-term horizons. Tighter external financial conditions have a marked negative effect on the entire distribution of the growth outlook over the near-term, with the negative impact more important than any of the other dimensions of financial risk. Weaker external demand also has a strong negative effect on the near-term growth outlook. This finding is consistent with the fact that Panama is a highly open economy with its business model founded on its ability to be an attractive destination for international financial, business and transportation services.8 The negative effect of both categories of external conditions dissipates as the forecast horizon lengthens.

7. On aggregate, the conditional information from the price of risk, leverage and external financial conditions is consistent with lower risks to the growth outlook from financial conditions in the near-term relative to the medium-term. The conditional information from the quantile regressions and the various dimensions of financial conditions is used to derive the probability distribution for Panama’s growth in 2018 (one-year ahead), 2019 (two-year ahead), and 2020 (three-year ahead).9 The distributions are calibrated so that the mode, or most likely outcome, is consistent with the forecast for Panama (i.e. 4.3 percent in 2018, 6.3 percent in 2019 and 5.8 percent in 2020). These density forecasts can then be used to estimate the growth at risk (GaR) associated with various states of the financial system. Given current financial conditions, the GaR model forecasts that under a severely adverse growth scenario (one with 5 percent probability) for 2018 the growth outlook for Panama would be 3.9 percent compared to the outlook for 4.3 percent growth.10 This compares to a severely adverse scenario of 1.4 and 2.3 compared to the outlook for 6.3 and 5.8 percent growth for 2019 and 2020, respectively.

uA03fig03

Panama: Growth Density Forecasts

(Probability)

Citation: IMF Staff Country Reports 2019, 012; 10.5089/9781484394069.002.A003

Sources: IMF staff calculations. Forecast densities for growth, one, two and three years ahead.

8. The distributions can also be used to assess the cumulative likelihood of the growth scenarios used by Panama’s Superintendency of Banks (SBP) in its stress-test scenarios. The severe scenario used by the SBP assumes growth at end-2019 of 2.5 percent for Panama, broadly equivalent to a GaR scenario with 10 percent probability for 2019, whereas a more severe scenario consistent with a 5 percent probability would be appropriate to evaluate tail risks. Caution is also warranted as the stress tests results assume a relationship between economic activity and asset quality estimated over a limited time period (broadly since 2000) where economic activity has remained robust and asset quality has remained relatively stable, with low levels of non-performing loans (see related Selected Issues Paper).

D. Conclusions

9. Accommodative financial conditions support economic growth in the near-term, but can contribute to the build-up of financial imbalances overtime and put economic growth at risk. Results from the GaR model for Panama suggest that still broadly accommodative financial conditions in Panama should continue to support near-term growth prospects, but that the prolonged period of accommodative financial conditions can contribute to the build-up of financial vulnerabilities, putting at risk financial stability and growth over the medium-term. For Panama, near-term growth prospects should continue to be supported by a still accommodative price of risk, but the turning of the leverage cycle and ongoing tightening of external financial conditions bear close monitoring, particularly as tighter global financial conditions gradually lead to an increase in the price of risk in the context of Panama’s dollarized economy. A rapid deterioration in external financial conditions in particular could significantly worsen the outlook for neat-term growth, with a further tightening of leverage likely to have the largest medium-term growth impact. These results highlight the importance of remaining vigilant and building resilience to emerging financial risks. In the context of Panama’s dollarized economy with no central bank, macroprudential policy and crisis preparedness/management have an enhanced role to play in both mitigating and managing these risks.

Figure 2.
Figure 2.

Panama: Coefficients from Quantile Regressions of Financial Conditions on Future GDP Growth

Citation: IMF Staff Country Reports 2019, 012; 10.5089/9781484394069.002.A003

1

Prepared by Kimberly Beaton with thanks to Romain Lafarguette for sharing his code and his expertise and to Romain, Adrian Alter, and Alan Feng for helpful comments and suggestions.

2

See IMF (2017) for an introduction to GaR approach.

3

The indices are normalized around zero for the period for which they are estimated. Therefore, the indices provide an indication of the relative tightness/accommodativeness of financial conditions only for the time period for which they are estimated.

4

This is also a limitation to the inclusion of the credit-to-GDP gap.

5

See also IMF (2017), Annex 3.3 for a detailed description of the methodology used to estimate the conditional density of future GDP growth based on current financial conditions.

6

For instance, IMF (2017) finds that rising funding costs and falling asset prices are the most important signals of severe recession at time horizons of up to four quarters.

7

Investment is estimated to have contributed on average 4 percentage points to Panama’s annual economic growth over 2008–2016 with a substantial share of this related to large scale infrastructure projects like the Panama Canal expansion (U.S. $5.3 billion over 2007–16) and a new mine (U.S. $5.5 billion). Many of these large infrastructure projects were financed externally, either through sovereign issuance on global capital markets or private capital inflows, largely concentrated in FDI rather than through the Panamanian banking system or capital markets.

8

See IMF Country Report No. 17/106.

9

Based on the conditional information from the quantile regressions, a t-skew fitted curve approach is used to derive the conditional distribution for Panama’s GDP growth for each of the forecast horizons considered. See IMF (2017), Annex 3.3 for a detailed description of the methodology.

10

For 2018, the assumption that the mode of the growth density forecast is 4.6 percent skews the distribution to the left. This forecast takes into account the impact of the abrupt stoppage in construction activity during the strike, which is estimated to reduce growth by 1 percentage point for 2018. Allowing the mode of the distribution instead to be consistent with the conditional mean suggested by the quantile regression results would give a forecast mode of 5.6, consistent with an estimated impact of the strike on the mode outlook for growth of 1 percentage point.

Panama: Selected Issues
Author: International Monetary Fund. Western Hemisphere Dept.