Abstract
Malta: Selected Issues
A Financial Conditions Index for Malta1
This paper develops a Financial Conditions Index (FCI) for Malta using a Principal Components Analysis (PCA). The constructed FCI shows an improvement in financial conditions in the post-crisis period on the back of more favorable domestic and external conditions. In addition, the FCI, which was found to have a high correlation with future economic activity and a significant impact on the dynamics of household credit, suggests that the current favorable financial conditions are likely to support economic growth in the near term.
A. Introduction
1. A Financial Conditions Index (FCI) is a useful financial surveillance and forecasting tool. It is often used as a financial surveillance tool as it conveniently summarizes the bulk of the informational content of an array of financial and economic indicators.2 As financial conditions are known to have an important influence on business cycles because they reflect markets’ expectations regarding future economic developments, an FCI can be used to improve the predictive power of econometric models for GDP growth and other economic indicators.3
2. In the absence of theoretical guidance, both judgment and parametric estimation are used to construct an FCI. Judgment is widely used to select the components of an FCI and assign their relative weights when the FCI is predominately used as a monitoring tool. A regression-based FCI and a principal component analysis (PCA) are the two parametric approaches that are commonly used in the literature. In the first approach, the weights of each financial indicator are assigned according to the estimated impact on real GDP growth in a vector autoregressive (VAR) or structural macroeconomic models. In the second approach, the FCI is based on a common factor, which is extracted from a group of financial indicators and captures the greatest common variation among them.
3. In this paper we use a PCA method to evaluate the financial conditions in Malta. The advantage of this method is its practicality as it allows one to quickly collapse a large set of financial variables into a single indicator. 4 However, as a purely statistical tool, a PCA has a major disadvantage because it assumes that the indicators with the greatest variability have the biggest economic significance. Therefore, and like any other approaches, the construction of the index requires judgement regarding the suitability of the indicators that are included.
B. Constructing a Financial Conditions Index for Malta
4. A PCA-based FCI for Malta is constructed based on asset prices, interest rates, exchange rates, and global risk indicators. We began with a fairly large data set that included various measures of consumer and asset prices, interest rates, exchange rates, risk and banking sector indicators, balance sheet indicators, and external sector indicators. We reduced the original set of quarterly indicators using the following criteria: (1) the time span should cover at least the last 10 years to capture the last two episodes of financial stress; (2) the sign corresponding to an indicator entering the first principle component has to be economically meaningful; (3) the relative weight of a standardized component in the FCI has to be greater than 10 percent.
5. The final set of FCI components includes external and domestic variables. Malta is a small open economy, and thus changes in the international environment can significantly affect financial conditions that the economy faces. Therefore, the FCI includes external indicators such as the Euro Area (EA) lending standards for non-financial corporates, the Ted-spread5 (which measures global credit risk), Chicago Board Options Exchange Volatility Index (VIX), and the Standard and Poor’s 500. Domestic variables that reflect Malta’s financial conditions are the real effective exchange (GDP deflator based), the 10-year sovereign yield spread with Germany, stock market prices, and housing prices (Central Bank of Malta’s index of advertised prices), and domestic banks’ capital and liquid assets ratios.6 All variables are demeaned, so the resulting FCI is centered around zero. All chosen indicators have a meaningful loading (of at least 40 percent).
Factor “Loadings”
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculationsFactor “Loadings”
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculationsFactor “Loadings”
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculationsC. Results
6. The results show that financial conditions have been favorable, particularly since end-2012. Financial conditions were quite supportive before the crisis. They started to deteriorate in 2007, mainly due to negative contribution from domestic variables (especially the stock market prices). As the international crisis unfolded in 2008, both external and domestic variables had a negative contribution to the FCI, but the contractionary impact turned positive in 2013 as both global and domestic conditions improved.7 Domestic factors were supported by the buoyancy in Malta’s stock market, the strong housing market momentum, the continued compression of the sovereign spread, and the REER dynamics. The improved external conditions also supported the financial conditions since 2012, though the increased financial volatility and the slower growth in US stock market prices posed some headwinds at end-2015.
Contributions to FCI
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Sources: IMF Staff CalculationsContributions to FCI
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Sources: IMF Staff CalculationsContributions to FCI
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Sources: IMF Staff Calculations7. The FCI is strongly correlated with GDP growth. Both visual inspection and simple correlations show that the FCI is highly correlated with future GDP growth (up to 5-6 quarters) over the entire sample. The high correlation between the FCI and GDP growth in 2006-14, is in part explained by large share of tradable sectors in the economy, which are susceptible to changes in external financial conditions and the exchange rate. The sizable financial sector and relatively high stock of private sector’s financial assets and liabilities also increase the sensitivity of the economy to changes in domestic and external financial conditions.
GDP Growth and FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculationsGDP Growth and FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculationsGDP Growth and FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculationsFCI Correlation with Real GDP Growth (y-o-y)
FCI Correlation with Real GDP Growth (y-o-y)
| t+1 | t+2 | t+3 | t+4 |
| 0.485 | 0.616 | 0.586 | 0.494 |
FCI Correlation with Real GDP Growth (y-o-y)
| t+1 | t+2 | t+3 | t+4 |
| 0.485 | 0.616 | 0.586 | 0.494 |
D. Forecasting Properties of the FCI
8. The high correlation of the FCI with real activity may suggest that it could serve as a useful tool to assess future real GDP developments. To address concerns about causality from economic activity to financial conditions, we regress the FCI on GDP growth to remove contemporaneous effects. The resulting residuals are used as a “purged” FCI. The trajectory of the purged FCI is broadly similar to that of the unpurged indicator. The estimations of the forecasting properties of the “purged” FCI show that it has meaningful predictive power. Econometric tests show that there is causality (in the sense of Granger (1988)) at 2 and 4 lags from the FCI to GDP growth, but not the other way around. Last, the forecasting power of a simple ARMA model of GDP growth is increased significantly by augmenting it with the FCI at different lags.
GDP Growth, Purged and Unpurged FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculationsGDP Growth, Purged and Unpurged FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculationsGDP Growth, Purged and Unpurged FCI
(percent)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: Haver, IMF Staff calculations
FCI Predictive Power
(increase in adjusted R2 relative to naive real GDP ARMA model)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculationsFCI Predictive Power
(increase in adjusted R2 relative to naive real GDP ARMA model)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculationsFCI Predictive Power
(increase in adjusted R2 relative to naive real GDP ARMA model)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff calculations9. Changes in financial conditions have a positive impact on credit. Estimates from a three-variable VAR with purged FCI, de-trended real credit to households, and real GDP growth show a positive and significant effect of the FCI on real credit and growth.8 In particular, the results suggest that an improvement in overall financial conditions have a direct effect on GDP growth within four quarters while contributing to the expansion of credit to households within two quarters.9 These results suggest that part of the positive effect of financial conditions on economic activity is taking place through the credit channel.
Pairwise Granger Causality Tests
Pairwise Granger Causality Tests
| Null Hypothesis: | P-value | |
|---|---|---|
| Lags | 2 | 4 |
| FCI does not Granger Cause RGDP | 0.003 | 0.028 |
| RGDP does not Granger Cause FCI | 0.819 | 0.463 |
Pairwise Granger Causality Tests
| Null Hypothesis: | P-value | |
|---|---|---|
| Lags | 2 | 4 |
| FCI does not Granger Cause RGDP | 0.003 | 0.028 |
| RGDP does not Granger Cause FCI | 0.819 | 0.463 |
Impulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff CalculationsImpulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff CalculationsImpulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff Calculations
Impulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff CalculationsImpulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff CalculationsImpulse Response Function
(Response to one standard deviation innovation in FCI)
Citation: IMF Staff Country Reports 2017, 057; 10.5089/9781475580082.002.A003
Source: IMF Staff CalculationsE. Conclusions
17. This paper develops an FCI for Malta. The analysis extracts the FCI using PCA approach, which identifies an unobserved common factor from a group of external and domestic financial indicators. The FCI indicates that the financial conditions have improved significantly since 2013 on the back of more favorable external and domestic conditions, though domestic factors have become more dominant recently. Moreover, the FCI, which was found to have a high correlation with future economic activity and a significant impact on the dynamics of household credit, suggests that the current favorable financial conditions are likely to support economic growth in the near term.
References
Angelopoulou, E., Balfoussia, H., and Gibson, H. (2013), “Building a Financial Conditions Index for the Euro Area and Selected Euro Area Countries: What Does It Tell Us About the Crisis?” European Central Bank Working Paper No. 1541.
English, W., Tsatsaronis, K., and Zoli, E. (2005), “Assessing the Predictive Power of Measures of Financial Conditions for Macroeconomic Variables,” BIS Papers No. 22.
Granger, C. (1988) “Some recent development in a concept of causality.” Journal of econometrics 39.1 (1988): 199–211.
Hatzius, J., Hooper, P., Mishkin, F., Schoenholtz, K., and Watson, M. (2010), “Financial Conditions Indexes: A Fresh Look after the Financial Crisis,” National Bureau of Economic Research Working Paper No. 16150.
Manning, J. and Shamloo, M. (2015), “A Financial Conditions Index for Greece”, IMF Working Paper No. 15/220.
Swiston, A. (2008), “A U.S. Financial Conditions Index: Putting Credit Where Credit is Due,” IMF Working Paper No. 08/161.
Prepared by Federico Grinberg.
See Manning and others (2015) as some examples of IMF surveillance done using FCI analysis.
English et al. (2005), Swiston (2008), and Hatzius et al. (2010) show that FCIs are highly correlated with GDP and have a strong predictive power for future economic activity.
The Ted-spread is constructed as the difference between the London interbank interest rate (Libor) and the US 3-month T-bill yield.
A wider range of domestic variables were examined; however, they were eventually dropped due to weak co-movement with the common factor or had the opposite sign that one would expect.
The Central Bank of Malta’s FCI also suggests that conditions have improved since 2013, though it shows that financial conditions remained less favorable compared to the pre-crisis period.
At a 5 percent significance level.
Results on overall credit to the private sector are inconclusive, potentially reflecting the recent non-financial corporates deleveraging.
