Does Country Size Matter? (Short Note)

Transcription

1 World Bank From the SelectedWorks of Mohammad Amin June 3, 2011 Does Country Size Matter? (Short Note) Mohammad Amin Available at:

2 Does Country Size Matter? (Short Note) Mohammad Amin Enterprise Analysis Unit, World Bank June, 2011 With the exception of trade openness, existing studies have failed to find any significant impact of country size on various social and economic variables. This note uses newly available firm-level and country-level data and shows that country size does matter with small countries performing better than large countries in areas such as trade facilitation, tax administration, burden of tax rates on private firms and corruption. The note also argues that the impact of country size on a variable of interest may not be uniform and it may depend on for example, how large the country is to begin with and its income level. According to the latest available estimates for 2009, there are over 65,000 Chinese for each of the 20,398 residents of Palau; there are 20 countries with a population of less than 100,000 and 11 countries with a population over 100 million. As these numbers suggest, differences in country size as measured by population are indeed large and breathtaking. Despite these large differences in country size, empirical work on whether country size matters or not for various social and economic variables is extremely limited. In fact, the few studies that do exist show that with the exception of international trade, country size does not matter much for various economic and social phenomena. For example, in a recent study covering over 200 countries and over 40 years, Rose (2006) fails to find any significant relationship between country size and a host of socio-economic variables such as the level of income, inflation, material wellbeing, health, education, the quality of a country s institutions and a number of other international indices and rankings. The only exception is international trade in that small countries are found to larger trade volumes (exports plus imports as a percentage of GDP) than large countries. This note takes another look at the relevance of country size using newly available firm-level and country-level data and it shows that country size matters for a number of important economic variables such as trade facilitation, tax administration, tax rates and corruption. 1

3 The irrelevance of country size in existing studies mentioned above is surprising because theory suggests a number of reasons why country size should matter. For example, Alesina and Spolaore (2003) list a number of country size related benefits that include lower per capita costs of public goods (monetary and financial institutions, judicial system, communication infrastructure, police and crime prevention, public health, etc.) and more efficient tax systems; cheaper per capita defense and military costs; greater productivity due to specialization (though access to international markets may reduce this effect); greater ability to provide regional insurance; and greater ability to redistribute income within the country. Against these potential benefits, there are some disadvantages of being large. For example, it is argued that larger countries have more diverse preferences, cultures and languages and this greater heterogeneity of preferences may make it more difficult to reach consensus on growth enhancing reforms. Another potential problem with large countries is congestion or administrative costs that may escalate with country size. The firm-level data we use are taken from a recent survey of over 11,000 non-agricultural firms in 30 countries in Eastern Europe and Central Asia conducted by the World Bank in The countries covered by the survey include Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, FYR Macedonia, Georgia, Hungary, Kazakhstan, Kosovo, Kyrgyz Republic, Latvia, Lithuania, Moldova, Mongolia, Montenegro, Poland, Romania, Russia, Serbia, Slovak Republic, Slovenia, Tajikistan, Turkey, Ukraine and Uzbekistan. The survey is designed to be representative of the non-agricultural private economy at the country level and contains a mixture of manufacturing and service firms, small and large firms and young and old firms. Country-level data are taken from the World Bank s Doing Business project and we use average values of the variables taken over all years for which the data are available ( ). Using the data mentioned above offers a number of advantages. First, these new data help to check how robust existing results in the literature are to alternative data sources. Second, the richness of our data allows us to test for new phenomena such as the quality of trade facilitation, tax administration and trade openness at the intensive as well as the extensive margin. Third, firm-level data allow us to take 2

4 account of differences across regions, sectors and firms within a country. Failure to control for these differences can potentially lead to misleading results, a worrisome issue for existing studies that are based solely on country-level data. Firm-level data confirm that small countries trade more than large countries It is argued that smallness of market limits the exploitation of economies of scale, forcing the relatively small countries to expand market size through international trade beyond their political borders (Alesina 2002, Alesina and Wacziarg 1998). Country level studies using the ratio of exports plus imports to GDP as a measure of trade openness confirm that small countries are more open that large countries (see, for example, Rose 2006). Do firm level data confirm that small countries are more open to trade than large countries? We find the answer to be in the affirmative (figure 1). i That is, the proportion of firms in a country that export and import is significantly higher in countries that are relatively smaller. Also, percentage of a typical firm s sales that is exported and the percentage of a typical firm s material inputs that are of foreign origin are significantly higher in the relatively smaller countries. We confirm that these findings are robust to various controls such as GDP per capita, firm-size, sector of activity and age of the firm. Figure 1: Exports as percentage of firm s sales is higher in smaller countries 3

5 Percntage of firm's annual sales exported EST SVN MKD LTU BIH CZE LVA HRV ALB SVK BLR KGZ HUN SRB MNGARMMDA TJKBGR GEO AZE KAZ ROM UZB POL UKR TUR RUS Population (log of average values over ) Source: Enterprise Surveys Small countries have better trade facilitation From the discussion above, it follows that trade is more important to small than large countries. If this is indeed true, one might expect small countries to invest more in ports and trade facilitation measures than large countries. In fact, trade facilitation measures can themselves be interpreted as alternative measures of trade openness. Hence, the question is whether trade openness and country size relationship mentioned above robust enough to extend to such trade facilitation based measures of trade openness. A recent study by Amin and Haidar (2011) puts the above question. The study uses an element of trade facilitation, the number of documents required to export and import (Doing Business data), as an alternative measure of trade openness and finds that small countries are indeed more open than large countries (figure 2). Specifically, a conservative estimate from the study shows that a 1 percentage point increase in population leads to 2.1 percentage point increase in the number of required documents to trade. This is an economically large and robust effect. Figure 2: Smaller countries have fewer documents required to trade 4

6 Number of documents required to trade (logs) KNA DMA GRD ATG SYC VCT VUT LCABLZ CAF RWAKAZ ERI AZE KGZ LAO BFA MWI CMR MRT COG KHM RUS NAMURY TJK AGO SYR NGA BTN SWZ BDI MLI GTMCIVNPL UGA MAR PRY NER ECU VEN DJI ALB SLV HNDSEN MOZDZAZAFTUR PAK MNG PNGIN BLR ZMB KEN SDN ARG ETH PHL GUYGABWA JOR TGOBOL TCD MDG LKA IRN VNM BRA LSO MKD GEOBEN BGR DOM MYS PERCOL UKR BGD MDA LBN CRI CHL GHA ROMTZA THA CPV MUS GNB JAM ARM LTU BIH NIC TUN EGY Population (log of average values over ) IDN IND CHN Source: Doing Business, World Bank. Note: The vertical axis uses average values over all years for which data are available ( ). Firms in small countries find tax administration more burdensome than firms in large countries Going beyond trade openness, Amin (2011a) looks at how the quality of tax administration as experienced by the firms varies between small and large countries. In one question, firms were asked to rank on a 0-4 scale how severe tax administration is as obstacle to their current operations. Higher values on the scale imply greater severity of the obstacle. One motivation here is that scale effects may lead more efficient tax administration in larger countries. However, congestion or administrative costs may also escalate with country size, implying less burdensome (to firms) tax administration. Findings from the paper tend to support the latter view. To cite one example, in one of the specifications, the paper finds that moving from the largest country to the smallest country in the sample increase the severity of the tax administration obstacle by about 71 percent of the average level of severity of the obstacle in the sample. Small countries have higher tax rates than large countries 5

7 The difference between small and large countries discussed in the previous section extends to tax rates too. This holds for both, the firm-level as well as country-level data. For example, using country-level data from Doing Business on the total corporate tax rate as a proportion of total profit ii, regression results show that according to the most conservative estimate, moving from the smallest to the largest country increases the tax rate by a large 79 percent of initial level of the tax rate. Figure 3 illustrates the point. Regression analysis also confirms that the stated country-size and tax rate relationship easily survives even when differences such as in income levels across countries are taken into account. For firm-level data, we use firm s response on 0-4 scale to tax rates as obstacles to their current operations, with higher values on the scale implying greater severity of the obstacle. Results using this measure also show a robust and large increase in the severity of the tax rates obstacle as country size increases. Figure 3: Corporate tax rate is higher in larger countries Total corporate tax rate (logs, average over ) PLW COM SLEBDI BLR ARG UZB MRT ERITJKBOL COL BEN YEMDZA ITA MHL COG BRA GNQ KGZ NIC TCD FRA FSM PRI TUN LKA URY CRI AUT HUN BEL ESP KNA CPV ALB AGO EST JAM CMR AUS ATG FIN CZE GRC GAB BFACIV BTN ARM AZE CAN EGY DEU GRD GNB PAN TGOSWE GIN MLI VEN UKR HND KEN MEX JPN SYC LTU ROM PHL RUS STP BHS SVK IRN DMA GUY DJI FJI GTM MDG LVABIH MDA NOR PNGPRT SEN VCT NER NLD SYRMAR TZA TUR LBR ZWE USA PRY BGR DOMKAZ PER POL SVN HTI TWN BLZ ECU AFG BGD IDN LCA SLB SWZ HRV GEO LAO SLV GHA KIR LBN NZL RWA MYS THA TTO MNG MOZ NPL UGA SDN VNM BRN ISRB ZAF GBR PAK JOR DNK KOR MNE MWI ETH NGA TON ISL SUR CHE IRL SGP IRQ HKG CHL CYPMUS LUX KSV LSO KHM OMN WSM TMP BWA MKDWBG ZMB BHR KWT SAU NAM ARE QAT MDV VUT Population (log of average value over ) CHN IND Source: Doing Business, World Bank. Note: The vertical axis uses average values over all years for which data are available ( ). Paying a bribe to get things done is more likely in larger countries 6

8 In the survey, firms were asked if it is common for firms in their line of business to have to pay some irregular additional payments or gifts to get things done with regard to customs, taxes, licenses, regulations, services, etc. Responses of firms were recorded on a 0-6 (Never to Always) with higher values implying more corruption. For the sample of manufacturing firms iii and using a large number of robustness checks and different estimation methods, regression results confirm a much higher level of corruption in large compared with small countries. Figure 4 illustrates the point. In terms of the magnitude and using the most conservative estimate, a move from the smallest to the largest country in the sample implies that the corruption index increases by 35 percent of the mean value of the index in the full sample. This is a large effect, economically and statistically. Figure 4: Corruption is more widespread in large than small countries Corruption index EST MKD SVN MNG LVA ALB ARM BIH KGZ TJK MDA SVK LTU GEO HRV AZE BGR BLR HUN CZE KAZ ROM UZB POL UKR TUR RUS Population (logs of average value over ) Source: Enterprise Surveys. Note: The corruption index on the vertical axis is measured on a 0-6 scale with higher values implying a firm more likely to pay a bribe to a government official to get things done (for more details, see the text above). The sample of firms in the figure is restricted to manufacturing firms alone. Some challenges and nuances in identifying the impact of country size 7

9 The analysis has so far been restricted to a simplistic relationship between a variable of interest and country size. For example, the analysis implicitly assumed that the magnitude of the impact of country size is same irrespective of the size of the country, its income level or other country and firm characteristics. Is it possible that country size, say through scale economies, may matter more for small than large countries or more for rich than poor countries? For instance, congestion or administrative costs may increase as a country becomes larger. However, it is possible that this increase may be less for the relatively richer countries that are able to devote more resources than the poor countries to limit the increase in congestion and administrative costs. In short, the relationship between country size and economic variables may be non-linear. To cite one example, Amin (2011b) looks at the time it takes to obtain the necessary documents to trade (time cost) and how this time cost varies with the number of documents required to trade. The data source for the variables is Doing Business. One hypothesis that being more reliant on trade due to their limited markets, small countries are likely to have more efficient customs and related trade facilitation system. Hence, an increase in the number of required documents may add less to the time cost in small compared with large countries. Using panel data at the country level, the study finds that the data do not reject the hypothesis. Similarly, the study by Amin and Haidar (2011) discussed above shows that while the number of documents required to trade increase with country size, the increase is much bigger for the relatively smaller countries. The presence of such non-linear effects implies nuances in how country size matters and a challenge for future research in properly identifying the importance of country size. Newly available firm-level and country-level data show that country size matters for the quality of trade facilitation, tax administration, tax rates and corruption. Without exception, small countries tend to perform better than the large countries in terms of less burdensome tax administration and tax rates to the private agents, less corruption and better trade facilitation. These results are in sharp contrast to the 8

10 existing literature that has failed to find any significant impact of country size on various socio-economic factors with the rare exception of trade openness. Our results suggest that using datasets such as firm level surveys that are richer in many dimensions that the ones typically used in cross-country studies and exploring more nuanced ways in which country size may affect socio-economic variables may help to better identify the importance of country size. References [1] Alesina, Alberto and Romain Wacziarg (1998), Openness, Country Size and Government, Journal of Public Economics, 69(3): [2] Alesina, Alberto (2002), The Size of Countries: Does It Matter? Harvard Institute of Economic Research, Discussion Paper No. 1975, Harvard University, Cambridge, Massachusetts. [3] Alesina, A. and E. Spolaore (2003), The Size of Nations, MIT Press, Cambridge. [4] Amin, M and J. I. Haidar (2011), Trade Facilitation and Country Size, Mimeograph. Paper is available here. [5] Amin, M (2011a), Quality of Tax Administration: How Relevant is Country Size? Mimeograph. Paper is available here. [6] Amin, M (2011b), The Time Cost of Documents to Trade, Mimeograph. Paper is available here. [7] Rose, Andrew (2006), Size Really Doesn t Matter: In Search for a National Scale Effect, Journal of Japanese International Economies, 20(4): i The analysis in this section is restricted to the sample of manufacturing firms alone. The restriction most closely approximates the broader literature that deals with merchandise trade alone. We do not find a significant effect of country size on trade for service sector firms covered by the survey. ii The total tax rate measures the amount of taxes and mandatory contributions borne by the business in the second year of operation, expressed as a share of commercial profit. A detailed definition of the variable is available here. iii Regression results for the sample of service sector firms do not show any significant impact of country size on reported corruption levels. That is, country size and corruption relationship is sector specific, an issue that we do not explore in this note for lack of space. 9