CC BY
19
Макарян А Р
роль импорта и кредитов в объяснении макроэкономической уязвимости армении
МАКАРЯН А.Р.
РОЛЬ ИМПОРТА И КРЕДИТОВ В ОБЪЯСНЕНИИ МАКРОЭКОНОМИЧЕСКОЙ УЯЗВИМОСТИ АРМЕНИИ
Аннотация. Оценивая воздействие реальных расходов на конечное потребление домашних хозяйств, государства, валового накопления основного капитала, и реального экспорта товаров и услуг на реальный импорт в течение двух разных периодов: с первого квартала 2000 года по второй квартал 2008 года, а также с первого квартала 2000 года по четвертый квартал 2014 года, автор выявляет те расходные компоненты ВВП, рост которых в основном повлечет за собой рост импорта в среднесрочной перспективе. Автор приходит к выводу, что рост реального импорта товаров и услуг будет в основном определяться изменениями валового накопление основного капитала, а также реального экспорта. Тем не менее, если произойдет импортозамещение в случае формирования валового накопления основного капитала с использованием местных товаров и услуг, и увеличение добавленной стоимости экспортируемых товаров создаваемых на территории страны с использованием импортируемого сырья и продуктов, это повлечет за собой рост импорта, но это не приведет к дальнейшему серьёзному ухудшению дефицита текущего счета. И путем оценки влияния кредитов частным компаниям и домашним хозяйствам, предоставляемые коммерческими банками за период 2005:1-2014:3, автор приходит к выводу, что экономический спад, вызванный внешними шоками, приведет к пересмотру политики банковского кредитования, который может негативно повлиять на реальный сектор экономики.
Ключевые слова: Импорт, экспорт, расходы на конечное потребление государства, расходы на конечное потребление домашних хозяйств, валовое накопление основного капитала, кредиты, дефицит счета текущих операций, макроэкономическая уязвимость
MAKARYAN АЯ.
THE ROLE OF IMPORTS AND CREDITS IN EXPLAINING THE MACROECONOMIC VULNERABILITY OF ARMENIA
Abstract. By estimating the impact of real household and government final consumption expenditure, real gross fixed capital formation and real exports of goods and services on the real imports for two different periods: from the first quarter of 2000 to the second quarter of 2008, and from the first quarter of 2000 to the fourth quarter of 2014, the author identifies which GDP expenditure component would mainly drive the growth of imports in the medium-term. The author concludes that growth of real imports of goods and services would be mainly be driven by the changes in gross fixed capital formation, and real exports. However, with import substitution taking place in case of gross fixed capital formation by relying more on domestically produced goods and services, and increase of valued added of locally produced items to be exported with use of imported inputs, would lead to increase in imports, but this won't result in further serious worsening of the current account deficit in the medium-term. And by estimating the impact of loans to private companies and households provided by commercial banks for the period 2005:12014:3, the author concludes that economic downturn caused by external shocks would lead to revision of bank loan policy, which would negatively affect the real sector of economy.
Keywords: Imports, exports, government final consumption expenditure, household final consumption, gross fixed capital formation, loans, current account deficit, macroeconomic vulnerability
Introduction
From 2003 till 2007 Armenia was reporting a double digit economic growth, while in 2008 the growth slowed down, and in 2009 the economy was hit by the Global Financial Crisis that turned into an economic and finical crisis. In 2009, a currency crisis was reported. Could the crisis have been predicted and/or forecasted by at least relying on the best indicator-predictors of the crisis? According to Lane and Milesi-Ferretti (2010) during the Global Financial crisis that hit the economies of nearly all nations of the world the "growth during the crisis was lower in countries with higher income per capita, high pre-crisis credit growth, and current account deficits." [1, p.13] Based on empirical findings Goldstein and et al. (2000) state that two low-frequency (annual) indicators, which could be used to forecast that a currency crisis would come out in emerging markets, were large current account deficit-to-GDP, and large current account deficit-to-investment ratios. [2, p.96] And based on the literature review Claessens and Kose (2013) state that "crises are typically preceded by somewhat larger current account deficits relative to historical averages, although credit trends more
than external imbalances appear to be the best predictor." [3, p.33] Therefore, two indicators that could be considered as a good predictor of a crisis (especially currency) are large current account deficit-to-GDP ratio, and the credit trends. Makaryan (2015) stresses that although the current account deficit-to-GDP ratio was a good predictor of the currency crisis that was reported in Armenia in 2009, however, due to contagion the country could experience the aftermaths of a currency crisis even when the ratio is rather low and doesn't exceed 9% (the threshold values when exceeded, economic and currency crises could be expected and/or forecasted according to Avetisyan and et al. (2012) [4, p.82]) [5, pp.76-77]. The large current account deficit-to-GDP (exceeding 10% from 2010 to 2014)1 ratio was mainly driven by changes in volumes of imported goods and services, while the balance of primary and secondary income accounts was positive and was explained by the growths of personal transfer and compensation of employees.
Total loans of banking system-to-GDP ratio reached 42.7%2 in 2014, while this ratio comprised only 33.6% in 2011. This trend indicates that Armenian private companies and households would face difficulties in repaying loans in case of economic downturn over the short-run and in medium-term, and especially if the loans were provided in foreign currency, hence increasing the number of non-performing loans. And stricter terms and conditions applied with regard to collateral and new loan applications received could result in credit crunch in case of sound business plans submitted by private companies, in particular. These negative developments would reduce the amount of loans (in terms of quantity and volumes) provided to the private sector, thus directly affecting the domestic demand and prolonging the period of recession and making the recovery process more slowly.
Therefore we attempt to estimate statistically significant changes in real imports of goods and services caused by changes in the main GDP expenditure components (namely, gross fixed capital formation, household final consumption expenditure, government final consumption expenditure, and exports of goods and services (real values)), and statistically significant changes in imports of goods and services caused by changes in volumes of loans provided by commercial banks to households and private companies. These estimates would allow the government to design proper crisis response measures in case Armenia faces an economic downturn caused by external shocks, and chose a respective strategic option for reducing the current account deficit.
Model Specifications
Our first model is defined as:
Real Imports of goods and services = f (real household final consumption expenditure; real government final consumption expenditure; real gross fixed capital formation; real exports of goods and services)
The dataset contains 60 observations covering the period 2000:1-2014:43. The same equation was estimated 2 times for 2 different sample periods in order to identify the shifts in the behavior of main economic agents, and components that drive the growth of imports, other things being equal before the Global Financial Crisis and afterwards. The sample period of the first estimation covers from the first quarter of 2000 to the second quarter of 2008, and the second estimation sample period is from the first quarter of 2000 to the fourth quarter of 2014.
All variables are in real terms, and seasonally adjusted by moving average method. In order to fix the problem of stationarity, the first difference is taken. And by testing for the evidence of serial correlation respective orders of MA and AR processes are included in the equations to fix the problem. Upon testing for the existence of multicolinearity and finding no evidence of it, the following equation is estimated (see table 1 and 2 for estimation output):
DRIMPORTSAt=p0+PiDRGFCFSAt+p2DREXPGOVSAt+ p3DREXPORTSAt+
+P4DREXHHSAt + Et (1),
1 Source: Ministry of Finance of the Republic of Armenia, The main macroeconomic quarterly indicators of the Republic of Armenia online database (http://www.minfin.am/index.php?cat=69&lang=3)
2 Source: Central Bank of Armenia, Statistics of financial organizations, Total assets of banking system/GDP/total loans of banking system/GDP online database (https://www.cba.am/en/sitepages/statfinorg.aspx).
3 Source: National Statistical Service of Armenia (NSS), Armenia's National Accounts 2009, 2010, 2011, 2012 (Annual and quarterly data), 2009, 2010, 2011, 2012, respective pages: 240, 212, 212, 214-215, NSS: Yerevan, Armenia (in Armenian). NSS, Expenditure of GDP (by main consumption elements), quarterly data online database (http://www.armstat.am/en/? nid=263), based on SNA 1993. Author's own calculations. Note: 2005=100. The values are in Armenian Drams.
Макарян А Р
_роль импорта и кредитов в объяснении макроэкономической уязвимости Армении_
where:
DRIMPORTSAt is the first difference of seasonally adjusted value of real imports of goods and services in period t.
DRGFCFSAt is the first difference of seasonally adjusted value of real gross fixed capital formation in period t.
DREXPGOVSAt is the first difference of seasonally adjusted value of real government final consumption expenditure in period t.
DREXPORTSAt is the first difference of seasonally adjusted value of real exports of goods and services in period t.
DREXHHSAt is the first difference of seasonally adjusted value of real household final consumption expenditure in period t.
Po, P1, P2, - • •, P4 are model unknown parameters. Et is the error term in period t. Our second model is defined as:
Real GDP = f (volume of credits provided to households and private companies by commercial banks)
The dataset contains 39 observations covering the period 2005:1-2014:3.4 We deliberately exclude the last quarter of 2014 since the value of volumes of loans is identified as an outlier. The same transformations are done here as well. The values of volume of loans for each period are in nominal terms, since owing to the lack of information regarding how households and private companies spent the loans on; it is hard to identify which price index would be applicable for the respective share of loans. We estimate the following equation (see table 3 for estimation output): DRGDPSAt= a 0+ a1 DLOANTt+ Et (2), where:
DRGDPSAt is the first difference of seasonally adjusted value of real GDP in period t. DLOANTt is the first difference of seasonally adjusted value of volume of credits provided to households and private companies by commercial banks in period t. a 0, a 1 are model unknown parameters. Et is the error term in period t.
Estimation Results
Imports Growth Prospects
Estimation Results state that before the Global Financial Crisis and for the period covering the entire dataset (Estimation #1, 2, Table 1, 2) the statistically significant changes in real imports of goods and services could be explained by changes in above discussed main GDP expenditure components (other things being equal), however, the time period (for some of them) and the value of coefficient is different. Based on two estimation results the followings could be concluded and expected (if migration growth rates are not hiking and remain unchanged):
1. The value-added of goods produced in Armenia and services rendered with use of imported inputs (both goods and services) and exported has grown. This means that changes in the amount of imports depend mainly on the fluctuations in volumes and structure of goods and services being exported. If growth of exports originates from increase of exports in low-tech, labor-intensive sectors, then imports will grow more than if the increase roots in exports of produce made in science-intensive, and medium-high-technology and high-tech sectors. Meantime, relying on growth rates of real spending on goods and services by tourists in Armenia from 2010 to 20145, increase in real export volumes due to higher numbers of tourists will be accompanied with certain changes in real imports of goods. Therefore, the priorities of the Government could be the followings:
- In case the current structure of exports remains unchanged, to promote growth of exports towards developed countries (specifically, EU) in the medium run to ensure higher value-added of
4 Source: Ibid., and Monetary and Financial Statistics of the Central Bank of Armenia, Commercial Banks Loans by Sectors (monthly) online database (https://www.cba.am/en/SitePages/statmonetaryfinancial.aspx). Author's own calculations. Note: the values are in Armenian Drams (AMD).
5 Source: Balance of Payments 6 online database of External account (http://www.armstat.am/am/?nid=443), Exchange rates by indicators and years online database of Foreign Trade (www.armstatbank.am), Prices and Price Indices in the Republic Armenia 2006-2010, 2007-2011, 2009-2013, 201-2014, 2011, 2012, 2014, 2015 respective pages: 14, 26,10, 10, NSS: Yerevan Armenia. Author's own calculations. Note: 2005=100.
items produced in Armenia, as well as to focus on producing goods with higher value-added (in the frames of the given sectors and structure).
- Along with development of science-intensive, and hi-tech sectors (to be strengthen in the short run or in the medium run), promote growth of exports volumes of these sectors in the long run.
2. Accounting for the circumstance of growing final consumption of households in 2010 to 2014, despite that growth did not exceed 1% from 2013 to 2014 as compared to the previous years6, and while growth of household real consumption per capita in 2011 to 2014 (excluding year 2013)7 was not accompanied with increase in household real consumption of food per capita from 2011 to 20148, we can conclude that changes in and timing of the average expected magnitude of statistically significant impact of changes in household final consumption (as compared to Estimation #1) reflected mainly the shift in the demands of households (towards consumption of non-food items and of services). In general, the growth rates of household final consumption from 2013 to 2014 evidence that if no major increase in household real income follows (accounting also for loans granted to households by commercial banks in Armenia9, the final consumption amounts will either stay roughly unchanged or will demonstrate a slight increase (in case that the Armenian GDP growth rates are at least the same). Therefore, growth in the volumes of this particular expenditure component in the medium run will not lead to a sharp change in real imports, while its reduction (depending on developments in the Armenian economy) will result in a decrease in real imports of goods and services.
3. According to the growth pattern of final consumption expenditure of public institutions from 2010 to 201410, as well as the Government's real spending of goods and services for the same period11 we can assume that there has been some sort of import substitution (more locally produced goods became available). If in the medium term the Government final consumption is accompanied by some increase of import substitution, the real volume of imports would not record a sharp rise. As a result of the economic decline in the medium term the real growth of spending on items and services as it happened in 200912 would lead to the increase in real imports of goods and services (other things being equal).
4. The changes in gross fixed capital formation were accompanied with an increase in the consumption of domestically produced goods and/or services. By taking into account the growth rates of gross fixed capital formation rates of 2010-2014, and namely the rates of private sector during the same period13, the statistically significant changes of real imports in the medium term will depend on the fact whether Armenia will attract "efficiency-seeking" foreign companies and/or investors that aim to enter the regional (Eurasian Economic Union, Iran) market, as well as companies that want to set up subsidiaries and/or branches in Armenia for serving the outsourced contracts of the head offices (especially from the EU) by benefiting from productivity-adjusted relatively low remuneration for labor, as well as by the way how would be changed the export-structure (depending on the economic growth in export destination markets) and the volumes exported, and what will be the competition in various sectors of the economy. Otherwise the declining investments spending by private sector could lead to a decline in real imports.
Therefore, in order to sustain and reduce the current account deficit (if primary and secondary income accounts remain unchanged that in it turn would depend on the economic situation in Russia) the Government of Armenia choose one (or both) of the below described strategic options in the mediumrun:
6 Source: NSS, Expenditure of GDP (by main consumption elements), quarterly data online database (http:// www. armstat. am/en/?nid=263)
7 Source: NNS, Social Snapshot and Poverty in Armenia statistical-analytical report, 2009, 2012, 2013, 2014, 2015, respective pages: 96, 113, 125, 113, 116, Prices and Price Indices in the Republic Armenia 2006-2010, 2007-2011, 2009-2013, 201-2014, 2011, 2012, 2014, 2015 respective pages: 14, 26,10, 10, NSS: Yerevan Armenia. Author's own calculations. Note: 2005=100.
8 Ibid.
9 Source: Monetary and Financial Statistics of the Central Bank of Armenia, Commercial Banks Loans by Sectors (monthly) online database (https://www.cba.am/en/SitePages/statmonetaryfinancial.aspx).
10 Source: NSS, Expenditure of GDP (by main consumption elements), quarterly data online database (http:// www.armstat.am/en/?nid=263)
11 Source: Prices and Price Indices in the Republic Armenia 2006-2010, 2007-2011, 2009-2013, 201-2014, 2011, 2012, 2014, 2015 respective pages: 14, 26,10, 10, NSS: Yerevan Armenia. The Ministry of Finance of Armenia, Macroeconomic indicators, The indicators of state budget of the Republic of Armenia (2008-2015) online database online Database (http:// www.minfin.am/index.php?cat=69&lang=3) Author's own calculations. Note: 2005=100.
12 Ibid.
13 Source: WB, WDI-Online Database (http://www.worldbank.org).
Макарян А Р
_роль импорта и кредитов в объяснении макроэкономической уязвимости Армении_
1. By relying on export promotion programs and measures the Government could ensure increase in exports volumes by choosing the first direction discussed above, in particular, that could be accompanied with the attraction of FDI that would enable to sustain the stability of the foreign exchange rate and ensure servicing of the foreign debt;
2. Targeting and attracting those foreign companies that could be classified as efficiency- and resource-seeking investors in order to sustain the current account deficit and report increase in volumes exported that would lead into reduction of the currency crisis.
Loans and Economic Growth
Based on Estimation #3 (Table 3) results it could be expected that decline in the amount of loans provided to both private companies and households by commercial banks changes by 1 billion AMD in period t, would, on average, cause GDP decline by 494.2 million AMD in the same period. These results indicate that that if Armenia faces an external shock and the commercial banks are be forced to apply stricter rules for selecting credit applicants it would negatively affect the real sector of economy (via "domestic demand channel") thus, slowing the process of recovery of the economy. Therefore, one of the priorities of the government in times of economic downturn is to subsidy credit interest rates in cases when loan applications received are aimed at acquisition of items to be considered as gross fixed capital formation. The recession could result in increase of number of non-performing loans, thus making commercial banks somehow vulnerable to developments in the real sector of economy owing to external shocks that could lead to banking crisis in the worst scenario.
Conclusions
The increase of the current account deficit driven by growth of imports makes the Armenian economy vulnerable to external shocks. Two expenditure components that would mainly explain statistically significant changes of the latter one are investments and exports in times when the economy is reporting growth. However it would not lead to strong increase in imports since the value added of exported goods produced locally with the use of imported inputs increased, and imports substitution is taking place in case formation of gross fixed capital, therefore growth of imports would not cause substantial changes in current account balance. And if attracted investments result in increase of exports it would reduce the current account deficit. In times of a crisis explained by the external shocks, the decline of the discussed four expenditure components would cause statistically significant decline in imports. The external shock that could lead to economic downturn could enforce the banks to revise their loan policy, thus negatively affecting the real GDP.
Table. Estimation output №1
Dependent Variable: DRIMPORTSA
Method: Least Squares
Sample(adjusted): 2001:1 2008:2
Included observations: 30 after adjusting endpoints
Convergence achieved after 11 iterations
Backcast: 2000:4
Variable Coefficient Std. Error t-Statistic Prob.
C -570.0699 1796.350 -0.317349 0.7537
DRGFCFSA(-3) 0.358021 0.170831 2.095764 0.0468
DREXPGOVSA 0.937653 0.509926 1.838804 0.0784
DREXHHSA(-3) 0.224638 0.118361 1.897916 0.0698
DREXPORTSA 0.642860 0.170721 3.765552 0.0010
MA(1) -0.577542 0.171906 -3.359641 0.0026
R-squared 0.633015 Mean dependent var 5454.400
Adjusted R-squared 0.556560 S.D. dependent var 17701.67
S.E. of regression 11787.77 Akaike info criterion 21.76437
Sum squared resid 3.33E+09 Schwarz criterion 22.04461
Log likelihood -320.4655 F-statistic 8.279562
Durbin-Watson stat 2.087943 Prob(F-statistic) 0.000116
Inverted MA Roots .58
Table. Estimation output №2
Dependent Variable: DRIMPORTSA
Method: Least Squares
Sample(adjusted): 2001:2 2014:4
Included observations: 55 after adjusting endpoints
Convergence achieved after 8 iterations
Variable Coefficient Std. Error t-Statistic Prob.
C -1695.581 1698.402 -0.998339 0.3230
DRGFCFSA(-l) 0.320817 0.080199 4.000256 0.0002
DREXPGOVSA 0.705538 0.360594 1.956596 0.0561
DREXPORTSA 0.426715 0.123941 3.442890 0.0012
DREXHHSA(-2) 0.290140 0.119137 2.435340 0.0186
AR(2) -0.439875 0.138805 -3.169027 0.0026
R-squared 0.497070 Mean dependent var 2325.026
Adjusted R-squared 0.445751 S.D. dependent var 20894.69
S.E. of regression 15555.67 Akaike info criterion 22.24491
Sum squared resid 1.19E+10 Schwarz criterion 22.46389
Log likelihood -605.7349 F-statistic 9.685810
Durbin-Watson stat 2.279408 Prob(F-statistic) 0.000002
Table. Estimation Output №3
Dependent Variable: DRGDPSA
Method: Least Squares
Sample: 2005:1 2014:3
Included observations: 39
Convergence achieved after 10 iterations
Backcast: 2004:4
Variable Coefficient Std. Error t-Statistic Prob.
C 6.97E+09 3.96E+09 1.757929 0.0873
DLOANT 0.494218 0.248705 1.987166 0.0546
MA(1) -0.437855 0.151266 -2.894602 0.0064
R-squared 0.166047 Mean dependent var 7.23E+09
Adjusted R-squared 0.119717 S.D. dependent var 4.58E+10
S.E. of regression 4.30E+10 Akaike info criterion 51.87835
Sum squared resid 6.64E+22 Schwarz criterion 52.00631
Log likelihood -1008.628 F-statistic 3.583956
Durbin-Watson stat 1.914772 Prob(F-statistic) 0.038067
Inverted MA Roots .44
References:
1. Lane, Philip and Milesi-Ferretti, Gian Maria (2010). "The Cross-Country Incidence of the Global Crisis, " IMF Working Paper No. 10/171, The International Monetary Fund, Washington, D.C.: United States, 40p. (number of pages in PDF file). Available at: http://www.imf.org/external/pubs/ft/wp/2010/ wp10171.pdf (last accessed: March 12, 2015).
2. Goldstein, M., C. Reinhart, and G. Kaminsky, 2000, Assessing Financial Vulnerability: An Early Warning System for Emerging Markets, Institute for International Economics, Washington, D.C.: USA, 150p.
3. Claessens, Stijn and Kose, Ayhan M. (2013). "Financial Crises: Explanations, Types, and Implications" IMF Working Paper No. 13/28, International Monetary Fund, Washington, D.C.: United States, 66p. (number of pages in PDF file). Available at: http://www.imf.org/external/pubs/ft/wp/2013/wp1328.pdf (last accessed: March 10, 2015).
4. Avetisyan, H., Grigoryan, V., Dallakyan, A., Sargsyan, H. (2012). "Approaches to Assessing the Stability of the Current Account in Armenia" CBA Working Paper 08/12-01, Yerevan: Central Bank of Armenia, 89p. Available at: https://www.cba.am/AM/panalyticalmaterialsresearches/2012_WP_1_CA_sustainability.pdf last accessed: March 19, 2015. (in Armenian)
5.Makaryan, A. (2015) "Current account deficit-to-GDP ratio as an indicator of Macroeconomic vulnerability of the Republic of Armenia," Armenia: Economy and Finance, No. 3-4 (175-176), pp.75-78, (in Armenian)
