i
The influence of the financial crises and the exchange rate ceiling on the Swiss
export
University of Amsterdam
Bachelor Thesis, Academic year 2016-2017 Faculty of Economics and Business
Economics and Finance Tom van der Meij, 10553355 Dhr. Rutger Teulings MSc Number of words: 6214 Date: 30-01-2017
ii Statement of Originality
This document is written by Student Tom Petrus Eduardo van der Meij who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
iii Abstract
The common conception is that uncertainties have a negative influence on export. However, this study suggested that this is not the case for countries that are seen as a safe haven, Switzerland is used to test this concept. In this paper an increase in uncertainties is formed by the mayor four financial crises in the last 36 years and a decrease is formed by an exchange rate ceiling, as introduced by the Swiss National Bank (SNB). If the effect of uncertainties on safe havens is indeed different, fiscal policy makers should take this in consideration to protect their countries export. The hypotheses are tested by using a panel data analysis over a period from 1980 until 2015, which contains the eighteen major importers of Swiss goods and services. The analysis results into two main findings. First, the occurrence the financial crises do not have an additional significant effect on the Swiss export. Secondly, the presence of an exchange rate ceiling does not lead to an additional significant effect on Swiss export.
iv Table of contents
Abstract ... iii
I. Introduction ... 1
II. Earlier research ... 2
III. Method ... 4
Methodology ... 4
Data ... 5
Data analysis ... 6
IV. Results ... 7
Pre- and Post-2007 Recession ... 7
Four Major Crises ... 9
GDP, Exchange Rate Ceiling and Exchange Rate ... 10
V. Conclusion, Discussion and further research ... 13
References ... 15
Appendix 1 ... 18
List 1 ... 18
Appendix 2 ... 19
1 I. Introduction
Over the past 36 years, several financial crises have struck the world trade. It is well know that during these events uncertainties arise around the globe. Recently Baldwin and Evenett (2011) have warned for these effects. They stated that these uncertainties would lead to a decrease in investing abroad and world trade and even inward turning of companies. In last few years, many papers have been written about these negative effects on international trade. However, these papers do not take into account that some countries are seen as a safe haven in times of uncertainty. One of these safe havens is Switzerland, a country that is known for its political, institutional, social and financial stability (Jordan, 2009). This leads to the question if these uncertainties have an additional significant effect on the Swiss trade. Hence new studies regarding the export of Switzerland are needed.
When taking uncertainties aside, the export of a country is based on the size of the foreign GDP (Kotan and Saygili, 1999) and the exchange rate between the country and its trading partners (Flemming, 1962). In this paper I will use two elements that have an influence on uncertainties. First the occurrence of a financial crisis, this leads to more uncertainties and the presence of an exchange rate ceiling, which leads to fewer uncertainties. Each of these two factors have a direct effect on foreign GDP and the exchange rate, that is why I look at their additional effect instead of their total effect on export. Analyzing Switzerland gives an opportunity to evaluate both the effect of an exchange rate ceiling and of financial crises on export for a safe haven. If additional effects are found, then fiscal policy makers need to take this into account when choosing their policy.
I focus on Switzerland and the eighteen biggest importers of Swiss goods and services, their GDP and the exchange rate for a period ranging from 1980 until 2015. I combine this data with the period of the exchange rate ceiling and the occurrence of the four biggest financial crises in the last 36 years. These crises are the Latin American Debt crisis, the Early 90s crisis the Internet Bubble and the 2007 Recession. After regressing the data I acquire the following results. First, making a distinction between pre- and post- 2007 Regression data does not lead to a significant difference between the independent variables of the Swiss export. Furthermore, the occurrence of one of the major financial crises in the last 36 years does not have a significant effect either. The same goes for the presence of an exchange rate ceiling during the time period. Finally, foreign GDP does have a significant effect and the exchange rate has only a significant effect when using lags. Based on these findings, it is concluded that financial crisis and the presence of an exchange rate do not provide significant effects on export that is not already identified by the foreign GDP and the exchange rate.
This paper is divided as followed. In the next section, previous researches are evaluated. In section III the methodology and data is discussed. In section IV the regression results including additional tests are elaborated. Section V provides a summary of this paper and an overall conclusion together with the discussion of my research with recommendations for future research.
2 II. Earlier research
In earlier research, a variety of studies related to variables that influence export and more specific the Swiss export can be found. In this section, I discuss the most relevant studies.
There are several variables that have effect on the export of a country. I will discuss the most common ones. Firstly the effect of the exchange rate, this effect is explained by the Mundel- Flemming model, based on the papers of Mundel (1963) and Flemming (1962). Mundel researched the theoretical and practical implications of the increased mobility of capital. Flemming looked at the different effects of monetary policy on floating exchange rate versus a fixed rate. The Mundel-Flemming model stated that a small open economy with perfect capital mobility will experience negative effect on export if there currency appreciated. A appreciation would lead the a relatively price increase of the goods in the home country compared to the foreign country. Which lead to a decreasing demand for products of the home country by foreigners and thus a decrease of the export.
Secondly there is the effect of foreign GDP. An increase of foreign GDP leads to an increase in foreign income followed by a growth in foreign demand. An example of this effect was shown by Kotan and Saygili (1999) when they researched the import demand of Turkey in the late 80s begin 90s. They wished to test the joint or individual significant of exchange rate, domestic demand and international reserves on the quarterly import. They found that the variables have a significant effect and that the exchange rate has the most effect on the short run while domestic demand and international reserves are the main determinants of import on the long run.
Thirdly and last are uncertainties that are caused by financial instability and crises. Investors tend to see uncertainties as risks and will be averse towards them unless they are compensated with a risk premium. If a sector, country or even international trade as a whole becomes more risky well the risk premium does not increase, investors are likely to look for alternatives. Baldwin and Evenett (2011) warned for the negative effects on export of such an event. They stated that the uncertainties caused by a crisis would lead to a decrease in investing abroad and companies would turn inward. They wrote their book with the intention to prevent such event by warning the G20 of these effects and there consequents.
The results of scientific research of these variables on the Swiss export have been similar when it comes to the effect of foreign GDP but papers are divided considering the exchange rate. As mentioned before, the 2007 Recession has led to an appreciation of the Swiss franc toward the euro.
3 According to Auer and Sauré (2011) does this appreciation not lead to damage to the Swiss export. They wrote that Swiss exporting companies are a special case because of their unique basket of exported goods compared to other exporting nations. This basket makes them only slightly sensitive to an appreciation of the Swiss franc and changes only slightly when the appreciation increases. They based their finding on historical data of the exchange rate and trade per sector. Grossmann, Lein and Schmidt (2016) have a similar paper that is also based on trade per sector covering a period from 1989Q1 to 2014Q4. They conclude that the Swiss export is relatively insensitive to changes in the exchange rate but is sensitive to the change in GDP of their importers.
In 2012 Auer and Sauré wrote a new paper in which they state that the appreciation of the Swiss franc at that time has led to a loss of revenue of 35 billion Swiss francs for the exporting companies. These loses are according to them only masked by the rise of the GDP of their importers. They based their new finding on export performance of the recent past to estimate the effect of exchange rate and demand growth in each export market. This is analyzed with both a regional and industrial dimension. The Swiss National Bank (2011) believed in the negative effect of an appreciation of the Swiss franc as well. Therefore they placed on 6 September 2011 until 15 January 2015 an exchange rate ceiling of 1.20 Swiss francs per euro. They stated that this was to protect the Swiss export.
Switzerland is not the first country to place limits on their exchange rate with the intention to improve export. Tenreyro (2007) disproves the positive effect of pegging a country’s currency to another on their trade flow. She states that countries that stop using a variable exchange rate and start pegging their coin with a large currency will not experience any significantly difference in their balance of trade. She bases her research on historical data from 100 countries ranging from 1970 until 1997 However, this paper together with Grossmann, Lein and Schidt (2016) and Auer and Sauré (2011) are contradicted by the fact that the Swiss export decreased with 2.6% in 2015 after the abolishment of the Swiss exchange rate ceiling. In this period the GDP of the Swiss importers kept on growing and the Swiss franc increased in value compared to the majority of its importers currency.
When it comes to risks, Switzerland is seen a safe haven in times of uncertainties. This is emphasized by Jordan (2009). He explains that this is because of Switzerland’s political ,institutional, social and financial stability ,low inflation, confidence in the central bank, comfortable official foreign reserves, high savings and net foreign asset position. This paper will analyze if the variable for export have a different effect on the safe haven Switzerland.
4 III. Method
To answer the research question, the eighteen major importers of Swiss products over a time period from 1980 until 2015 are analyzed by using a panel data regression. In this section, I elaborate on the methodology and I discuss the data sources.
Methodology
I use a panel data analysis to estimate the effect of the foreign GDP, the exchange rate, the occurrence of financial crises and the presence of an exchange rate ceiling on Swiss export. In the paper of Grossmann, Lein and Schmidt (2016) the variables have 3 dimensions. A time dimension t that indicates the quarter and year, a country dimension i that indicates the destination country and a sector dimension that indicates the sector that export the goods. Because this paper focuses on the effects of on the Swiss export as a whole, I do not distinguish between different individual sectors. The dependent variable of equation 1 and independent variable 0 through 5 form together the equation used by Grossmann, Lein and Schmidt (2016). They used a first different of export instead of export level because they wished to research the elasticities of Swiss exports across sectors and destination countries. For this paper I add the POST PRE dummies together with the exchange rate ceiling and EU dummy and remove the sector dimension this creates equation 1. This is used to test if the effects of the independent variables are significantly different after the 2007 Recession compared to before. The 2007 Recession is chosen because it is the largest crisis from the last 36 year and according to Baldwin (2009) changes global export significant.
( )
( ) , 1
Where t stands for the time period and i stands for the export destination. EXPit denotes the
volume of export from Switserland to country i at time t , S denotes the nominal exchange rate between Switzerland and the importer in Swiss franc, GDP is the nominal GDP country i in Swiss franc. POST and PRE are post and pre the 2007 Recession dates, respectively. PRE is 1 when the data is from 1980 until 2007 otherwise 0 and vice versa. CEIL indicates if an exchange rate ceiling is present. EU stands for membership of country of the European Union or the European Communities at time t. All variables (excluding dummy variables) are in logs and the independent variables S, EXP and GDP are lagged by one period.
5 When using a lacked dependent variable, a Nickell bias (1981) emerges. However, the bias is divided by T which is the number of conversations. Because the number of observations in this paper is large, the bias is insignificant. Normally a time fixed effect is added to account for the time effect on the dependent variable. However the variable for the exchange rate ceiling is only time dependent, this means that adding an additional time variable would lead to multicorrelation. A country fixed effect is used because changes in GDP and exchange rate are country dependent. The EU dummy is added to distinguish between EU countries and others. Being an EU member means less trade restriction which leads to a higher export.
The regression as presented by equation 1 only makes a distinction between the period before and after the 2007 Recession. However, this only shows the effect of the 2007 Recession on the independent variable. To test the direct effect of financial crises on the Swiss export, equation 2 is used. Over the period from 1980 until 2015, a number of financial crises occurred. The largest four are the Latin American Debt crisis, the Early 90s crisis, the Internet Bubble, and the 2007 Recession. To see the effect of each financial crisis individually, each crisis has its own specific dummy variable. This regression is given by equation 2.
2
Where LC, E90, IB and 07R are dummy crisis variables. These variables are set 1 in the time of their crisis and 0 otherwise. LC represents the Latin American Debt crisis from 1980 until 1983, E90, the Early 90s crisis from 1990 until 1993 and the IB, Internet Bubble from 2001 until 2003 and finally 07R, the 2007 Recession from 2008 until 2012.
Data
The analysis in this paper is based on yearly trade data from the Swiss Federal Customs Administration dated from 2015 back to 1980. This bureau provided the total nominal Swiss export and the export shares of the eighteen largest importers of Swiss goods and services in Swiss franc. Together these markets cover around 81% of the Swiss export. The countries are listed in appendix 1, ranked based on their average import since 1980. The yearly non-seasonally adjusted nominal GDP in American dollars at current price of the eighteen major importing countries is taken from the World Bank world development indicators. These are converted to Swiss franc with the nominal exchange rate. The nominal exchange rates of these eighteen countries are taken from the International Monetary Fund - international financial statistics and are based on yearly average market rates.
6 Dates related to EU memberships are acquired from the website europa.eu. The level of the exchange rate ceiling and the exact time period in which it was active are retrieved from SNB.CH. The World Bank website provided the data’s for the crisis. Definitions of the data as used in the regressions are shown in Table 1. The statistic summary can be found in appendix 2 together with the internal correlation between the variables.
Table 1. Variable Definitions
This table presents the variables that are included in the regressions and their definition.
Variable Definition
EXP the yearly export of Switzerland
S the yearly exchange rate of Switzerland and country i
GDP the yearly GDP of country i
PRE Prior to the 2007 Recession
POST Post to the 2007 Recession
EU Membership of European union or European Communities
CEIL Presence of an exchange rate ceiling
LC Occurring of the Latin American Debt crisis
E90 Occurring of the early 90s crisis
IB Occurring of Internet Bubble
07R Occurring of the 2007 Recession
Data analysis
Three potential problems can occur when using this methodology. First of all I have to test whether it is appropriate to use random effects (RE) or I have to use fixed effects (FE). Fixed effect allows the individual-specific effect to be correlated with the regressors. The random effect assumes that these effects are distributed independently of the repressors. Assuming the wrong effect will lead to biased and inconsistent estimations (Nakamura, 1981). Therefore, a Durbin-Wu-Hausman test is conducted to see if we have to use fixed effect or a random effect.
Secondly, the results can be influenced by heteroscedasticity between the variables. This means that the variable of the error remains not the same over a period of time. In case of heteroscedasticity the normal standard errors cannot be used. A Breusch–Pagan test (1979) is used to see if there is heteroscedasticity between the variables.
7 Third and finally, the results can be influenced by serial correlation. Serial correlation leads to an underestimation of the standard errors. A Wooldridge test is used to see if there is serial correlation. Serial correlation leads to biases of the standard errors and makes the results less efficient (Drukker 2003). If there is both heteroscedasticity and serial correlation, a robust term needs to be placed at the end of the regression to compensate for this.
The results of these three tests on the regressions can be found in appendix 3. The test concludes that all regressions need to use fixed effects, contain heteroscedasticity and serial correlation. To compensate for this, the regression will use a fixed effect and robust standard errors instead of the normal ones.
IV. Results
For this research, several panel data regressions have been conducted. In this section, I present the regression results.
Pre- and Post-2007 Recession
To test whether the 2007 crisis has a significant effect on the independent variables of the Swiss export, we use regression equation 1. The regression equation is used to conduct a two sided Wald t test with an alpha of 5% on the independent variables. On the dummies I conduct a one sided test because their value will be either 0 or 1 which means that they can’t be negative. If the t-value is bigger than the required value that comes with an alpha of 5%, the effect is. The regression results of equation 1 are presented in table 2.
The results of the first test indicate that none of the PRE variables are significantly different from their POST counterpart when using a significant level of 5%. This means that a distinction between post- and pre- the 2007 Recession has no significant effect the independent variable off export. Furthermore is the effect of the exchange rate ceiling insignificant this in contrast to the second hypothesis. The EU is significant but negative, this goes against the expectations. The values of the parameters for foreign GDP are similar to ones that were acquired by Grossmann, Lein and Schmidt in 2016. Both have a positive value but the effect of their values is much larger, 1.35 for the short run and 1.60 for the long run. The exchange rate is parameters are not similar because in this paper only the pre lagged exchange rate is significant while POST and PRE are not . See table 2
8 Table 2: Regression output 1.
This table shows the coefficients and robust standard errors of the first regression and the results of the Wald t test on the PRE and POST data
VARIABLES Coef. Robust Std. Err. Wald t test
PRE* -0.20*** (0.03) PRE* =POST* 0.77
PRE* 0.01** (0.01) PRE* =POST* 0.17
PRE* 0.22*** (0.03) PRE* =POST* 0.68
PRE* -0.01 (0.01) PRE* =POST* 0.27
PRE* 0.54*** (0.07) PRE* =POST* 0.71
POST* -0.19*** (0.01) POST* -0.01 (0.01) POST* 0.21*** (0.02) POST* -0.67 (0.58) POST* 0.65* (0.35) CEIL 0.04 (0.03) EU -0.01** (0.00) Constant -0.68 (0.16) Observations 630 R-squared 0.12
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
9 Four Major Crises
The first regression only looks at the difference between the independent variables of export before and after the 2007 Recession. To look at the effects financial crises on the export itself I added the four mayor crises as dummy variables in the regression to form regression 2. These dummy variables are LC, E90, IB and GR.
The new regression results are subjected to the Wald t test and the results are presented in Table 3. The effect of each dummy is smaller than the required value under alpha is 5%. This means that none of the crisis dummies has a significant additional effect on the export of Swiss goods and services to its biggest importers. This goose against the first hypothesis. Furthermore are the lagged export, lagged foreign GDP, first difference of Foreign GDP and EU significant with an alpha of 5%. The values of the parameters for foreign GDP are again similar to the ones that were acquired by Grossmann, Lein and Schmidt in 2016. Both have a positive value but the effect of their values is still larger. Like the first the regression, the exchange rate is parameters are not similar because in this paper they are not significant. When using an alpha of 10%, the dummy variables related to the Latin American Debt crisis, Internet Bubble, the 2007 Recession and the lagged exchange rate are significant. However, an alpha of 10% is non-standard and therefore in the rest of the regressions, the crisis dummy variables are omitted. The dummy for the Early 90s crisis and the variable for the first difference of the exchange rate were in both cases insignificant. This time the exchange rate ceiling is significant with an alpha of 5% and has a positive value, this is similar to hypothesis two.
10 Table 3: Regression output 2.
This table shows the Coefficients and Robust Standard errors of the second regression and the results of the Wald test on the crisis dummies
VARIABLES Coef. Robust Std. Err. Wald t test
-0.22*** (0.04) 0.01* (0.00) 0.20*** (0.03) -0.00 (0.01) 0.55*** (0.08) LC -0.02 (0.01) LC=0 0.11 E90 -0.00 (0.01) E90=0 0.77 IB 0.01 (0.01) IB=0 0.16 07R 0.03 (0.02) 07R=0 0.13 CEIL 0.03* (0.02) EU 0.02*** (0.01) Constant -0.31 (0.29) Observations 630 R-squared 0.11
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
GDP, Exchange Rate Ceiling and Exchange Rate
To test whether GDP, exchange rate and the exchange rate ceiling have a significant effect on the export of Swiss goods and services, I regress equation 2 without the crisis dummy variables.
When testing regression 3 with the Wald t test I acquire the results as presented in Table 4. These new regression results are tested by using the Wald t tests. The lagged export, both parameters for foreign GDP, lagged exchange rate and the EU dummy have a significant effect on export with an alpha of 5%. Like regression 2, the parameters of foreign GDP are positive but till much smaller than the parameters of Grossmann Lein and Schmidt. The value of the lagged exchange rate in the regression is significant and positive like their parameter but far smaller than their 0.77. The exchange rate ceiling is like in regression 1 insignificant which goose against hypothesis 2.
11 Table 4: Regression output 3.
This table shows the coefficients and robust standard errors of the third regression and the Wald test on the remaining variables.
VARIABLES Coef. Robust Std. Err. Wald t test
-0.20*** (0.03) = 0 0.00 0.01** (0.00) = 0 0.04 0.21*** (0.03) = 0 0.00 -0.00 (0.01) = 0 0.77 0.53*** (0.07) = 0 0.00 EU -0.01** (0.00) EU = 0 0.01 CEIL 0.018 (0.012) CEIL = 0 0.14 Constant -0.60*** (0.19) Observations 630 R-squared 0.11
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The regression results show the following findings. First of all, the change in export is negatively related with the export of previous year. Stated differently, when the export in year t-1 decreases with 1% relative to year t-2, the export in year t increases with 0.20% relative to year t-1.
Second, the exchange rate and GDP of previous year are positively related to the change in export of the upcoming year. The coefficients indicate that when GDP and exchange rate in year t-1 individually increases with 1% relative to year t-2, the export in year t increases with respectively 0.21% and 0.01% individually. The positive signs of both coefficients are in line with theory.
Third, the change in exchange rate is insignificant while the change in GDP is positively related with the change in export. When GDP over year t increases with 1% relative to year t-1, the export over that same time period increases with 0.53%.
Fourth and finally, being part of the European Union or the European Communities leads to a decrease in export in the current year relative to year t-1 of 0.01%. This goes against the expectations. Being a part of the European Union or European Communities should means less trade restriction and should lead to an increase in export. The deviation of the expectation could be explained by the Simpson’s Paradox (1951). This means that there is omitted variable bias which causes this paradox.
12 If the omitted variable has a positive coefficient and it is negatively correlated with the EU, the bias is negative. In that case, the size of the indirect effect is presumable bigger than the positive direct effect, causing a Simpson’s paradox. This lead to a negative coefficient of the EU well it should be positive. A second explanation could be that the EU dummy is only country specific and not time specific. When a country is going to be a member of these organizations, the market would know it years before it becomes true, which leads to an increase of export to these countries. This assumption makes the dummies only country dependent which leads to multicorrelation because a country FE is already used. This could also explain why the effect of the EU variable is different from the expectations.
13 V. Conclusion, Discussion and further research
In this paper I tested if the occurrence of a financial crisis and the presence of an exchange rate ceiling have an additional effect on the Swiss export. To answer this research question, the focus was placed on the export to the eighteen biggest importers of Swiss goods in a period from 1980 until 2015. I used an adjusted model of Grossmann, Lein and Schmidt (2016) which did not contain a sector specific data. Additionally, dummy variables were introduced that indicated the presence of a financial crisis occurring in that specific year and an exchange rate ceiling.
First of all, it was expected that the presence of a financial crisis would have an additional significant effect on the export of Switzerland. I found that the distinction between post and pre 2007 crisis does not lead to a significant effect on the independent variables of the change in Swiss export. Furthermore I found that the four mayor crises between 1980 and 2015 did not have a significant effect of change in Swiss export either. The second expectation was that the presence of an exchange rate ceiling would have an additional significant effect on the change in the Swiss export. In the first and the final regression it was discovered that the exchange rate ceiling form roughly 2012 until 2014 did not have a significant effect on the export of Switzerland as well. It had a significant positive effect in regression 2 but because regression 3 had lesser insignificant variable, it was seen as a better model so its results were used for this conclusion. Therefore has the second hypothesis been rejected as well.
A possible explanation for the rejection of both hypotheses is that Switzerland is seen by foreigners as a safe haven, which means that the import of Swiss goods and services is only influenced by exchange rate and their GDP and not by fear for uncertainty. If the importers do not fear the bankruptcy of Switzerland’s exporting firms, the crises do not have an additional effect on their import that is not already accounted for in their GDP. This might explain why the financial crises do not have an additional effect on the Swiss export. This can also explain the insignificance of the exchange rate ceiling. If the importers of Swiss goods and services see the finance stability in Switzerland as unquestionable, the ceiling would not provide any extra feeling of security. Thus having an exchange rate ceiling would not have a significant effect that is not already specified by the exchange rate variable.
14 The research contains several limitations, the first limitation is related to the frequency of the data. Because the data is annually, there are only three points in time where there was an exchange rate ceiling present. To get a better understanding of the effect of the exchange rate ceiling, quarterly or even monthly data should be used. This would lead to more data points. Also, it would be more specific because the ceiling was placed on a specific data and not the beginning of a year. For example, in September 2011 the exchange rate ceiling was set. However, in the regressions, the ceiling dummy was set equal to zero because the ceiling was absent for more than eight months. Using data with a higher frequency would lead to more precise dummy variable and therefore more precise results. Furthermore, in this study the focus was placed on Switzerland as a safe haven. However as stated before, Swiss exporting companies are seen as special because of their unique combination of exporting goods compared to other exporting nations (Auer and Saurè, 2011). For future research, it can be investigated whether the results as presented in this paper are also applicable to other safe haven countries.
15 References
Auer, R., & Sauré, P. (2012). CHF strength and Swiss export performance–evidence and outlook from a disaggregate analysis. Applied Economics Letters, 19(6), 521-531.
Auer, R., & Sauré, P. (2011). Export basket and the effects of exchange rates on exports–why Switzerland is special. Globalization and Monetary Policy Institute Working Paper, 77.
Baldwin, R. E. (2009). The great trade collapse: Causes, Consequences and Prospects. Retrieved from voxEU.org
Baldwin, R., & Evenett, S. (2011). The collapse of global trade, murky protectionism and the crisis:
Recommendations for the G20. Retrieved from voxEU.org
Breusch, T. S., & Pagan, A. R. (1979). A simple test for heteroscedasticity and random coefficient variation. Econometrica: Journal of the Econometric Society, 1287-1294.
Chen, S. (2012). The Implication of the Exchange Rate Floor in Current Times: The Swiss Experience. University of California, Berkeley.
Datastream International. (August 09, 2016). FCA - Federal Customs Administration, Switzerland [Online]. Available: Datastream International/Economics.
Datastream International. (August 09, 2016). IMF - International Financial Statistics [Online]. Available: Datastream International/Economics.
Datastream International. (August 09, 2016). World Bank - WDI [Online]. Available: Datastream International/Economics.
Decline at a High Level." The Federal Customs Administration. Foreign Trade Statistics, 01 Jan. 2016. Web. 09 Dec. 2016.
Drukker, D. M. (2003). Testing for serial correlation in linear panel-data models. Stata Journal, 3(2), 168-177.
16 Fleming, J. M. (1962). Domestic financial policies under fixed and under floating exchange rates.
Staff Papers, 9(3), 369-380.
Foulkes, Imogen. "Swiss National Bank Acts to Weaken Strong Franc." BBC News. BBC, 06 Sept. 2011. Web. 09 Aug. 2016.
Hanslin Grossmann, S., Lein, S. M., & Schmidt, C. (2016). Exchange rate and foreign GDP elasticities of Swiss exports across sectors and destination countries. Applied Economics, 1-17.
Jordan, T. J., Nationalbank, S., & der WGZ-Bank Luxembourg, S. A. (2009). Der Schweizer Franken
und die Finanzmarktkrise. SNB.
Kotan, Z., & Saygılı, M. (1999). Estimating an import function for Turkey. Central Bank of the Republic of Turkey, Research Department.
Mundell, R. A. (1963). Capital mobility and stabilization policy under fixed and flexible exchange rates. Canadian Journal of Economics and Political Science/Revue canadienne de economiques et
science politique, 29(04), 475-485.
Nakamura, A., & Nakamura, M. (1981). On the relationships among several specification error tests presented by Durbin, Wu, and Hausman. Econometrica: journal of the Econometric Society, 1583-1588.
Nickell, S. (1981). Biases in dynamic models with fixed effects. Econometrica: Journal of the
Econometric Society, 1417-1426.
"Publications." Publications. World Bank, Web. 09 Aug. 2016.
Siliverstovs, B. (2016). The franc shock and Swiss GDP: how long does it take to start feeling the pain?. Applied Economics, 48(36), 3432-3441.
Simpson, E. H. (1951). The interpretation of interaction in contingency tables. Journal of the Royal
17 Tenreyro, S. (2007). On the trade impact of nominal exchange rate volatility. Journal of Development
Economics, 82(2), 485-508.
18 Appendix 1
List 1
This list shows the average export share of the eighteen biggest importers of Swiss goods and services over the time period 1980-2015
Country Average export share 1980-2015
Germany 21.62 United States 9.90 France 8.79 Italy 8.26 United Kingdom 4.88 China 1.94 Japan 3.58 Hong Kong 2.55 Austria 3.36 Spain 2.90 Belgium 2.08 The Netherlands 2.93 Singapore 1.24 Canada 1.11 UAE 0.80 Saudi Arabia 0.97 South Korea 0.93 Australia 0.93
19 Appendix 2
This table shows the summary of statics of the 18 importing countries including the mean, standard deviation, minimum and maximum,. The variables are defined in table 1 and the sample period is from 1980 until 2015.
Table 5: Export in Swiss franc
Country Mean STD Min Max
Germany 2,53E+10 1,03E+10 9,57E+09 4,19E+10
United States 1,22E+10 6,80E+09 3,54E+09 2,74E+10
France 1,03E+10 3,93E+09 4,28E+09 1,77E+10
Italy 9,79E+09 4,44E+09 3,64E+09 1,82E+10
United Kingdom 5,98E+09 2,27E+09 1,18E+08 1,00E+10
China 2,62E+09 2,90E+09 2,33E+08 8,93E+09
Japan 4,16E+09 1,79E+09 1,26E+09 6,86E+09
Hong Kong 2,96E+09 1,83E+09 5,86E+08 7,02E+09
Austria 4,04E+09 1,43E+09 2,14E+09 6,32E+09
Spain 3,53E+09 2,20E+09 8,99E+08 7,51E+09
Belgium 2,55E+09 1,28E+09 1,13E+09 5,63E+09
The Netherlands 3,49E+09 1,62E+09 1,25E+09 6,19E+09 Singapore 1,48E+09 9,89E+08 2,62E+08 3,62E+09
Canada 1,43E+09 9,73E+08 4,18E+08 3,33E+09
UAE 1,05E+09 9,63E+08 1,70E+08 3,13E+09
Saudi Arabia 1,31E+09 4,72E+08 7,67E+08 2,84E+09 South Korea 1,09E+09 8,03E+08 1,11E+08 2,71E+09 Australia 1,15E+09 6,82E+08 3,01E+08 2,41E+09
20 Table 6: GDP in Swiss franc
Country Mean STD min max
Germany 2,90E+12 7,72E+11 1,57E+12 4,13E+12
United States 1,23E+13 3,99E+12 4,80E+12 1,79E+13
France 2,10E+12 5,97E+11 1,18E+12 3,20E+12
Italy 1,72E+12 5,05E+11 7,97E+11 2,64E+12
United Kingdom 2,03E+12 7,48E+11 9,47E+11 3,56E+12
China 2,67E+12 2,91E+12 3,18E+11 1,05E+13
Japan 5,00E+12 1,45E+12 1,82E+12 7,99E+12
Hong Kong 1,83E+11 7,98E+10 4,84E+10 2,98E+11
Austria 2,98E+11 9,92E+10 1,37E+11 4,64E+11
Spain 9,55E+11 4,23E+11 3,58E+11 1,78E+12
Belgium 3,66E+11 1,17E+11 1,84E+11 5,66E+11
The Netherlands 6,13E+11 2,19E+11 3,18E+11 1,01E+12 Singapore 1,30E+11 8,38E+10 1,99E+10 2,82E+11
Canada 1,10E+12 3,98E+11 4,59E+11 1,76E+12
UAE 1,68E+11 1,08E+11 5,31E+10 3,66E+11
Saudi Arabia 3,36E+11 1,74E+11 1,28E+11 6,91E+11 South Korea 7,22E+11 3,96E+11 1,14E+11 1,35E+12 Australia 6,93E+11 3,58E+11 2,51E+11 1,45E+12
21 Table 11: Exchange rate in Swiss franc.
Country Mean STD min max
Germany 1,13 3,20E-01 8,21E-01 1,64
United States 1,44 3,90E-01 8,88E-01 2,46
France 8,21E-01 6,02E-01 2,37E-01 1,64
Italy 6,78E-01 7,36E-01 7,26E-04 1,64
United Kingdom 2,37 6,37E-01 1,42 3,95
China 3,51 3,29E-01 1,37E-01 1,15
Japan 1,10E-02 1,75E-03 7,39E-03 1,57E-02
Hong Kong 1,94E-01 6,53E-02 1,14E-01 3,51E-01
Austria 7,41E-01 6,76E-01 1,17E-01 1,64
Spain 6,85E-01 7,29E-01 9,48E-03 1,64
Belgium 7,00E-01 7,15E-01 3,98E-02 1,64
The Netherlands 1,08 3,63E-01 7,31E-01 1,64 Singapore 8,39E-01 1,08E-01 7,00E-01 1,12
Canada 1,15 2,76E-01 7,52E-01 1,81
UAE 3,93E-01 1,06E-01 2,42E-01 6,69E-01
Saudi Arabia 3,93E-01 1,17E-01 2,37E-01 6,78E-01 South Korea 1,62E-03 6,78E-04 8,01E-04 2,92E-03
Australia 1,14E 4,03E-01 7,23E-01 2,26
Table 12: Correlation table
This table shows the correlation between the used variable. When a correlation has a star symbol indicated, then this correlation is significant for a 5% error
FDLEXP LLEXP1 LLS1 LLGDP1 FDLS FDLGDP FDLEXP 1 LLEXP1 -0.14* 1 LLS1 -0.06 0.21* 1 LLGDP1 -0.06 0.71* 0.02 1 FDLS 0.03 0.05 -0.12* 0.01 1 FDLGDP 0.27* -0.20* -0.09* -0.13* 0.16* 1
22 Appendix 3
Table 13: After regression test
This table shows the test that were performed after the regression in chapter 4 Hausmann test The Breusch-Pagan test Wooldridge test
Prob > Chi Prob > F Prob > F
Regression 1 0.00 0.00 0.00 Regression 2 0.01 0.00 0.00 Regression 3 0.01 0.00 0.00 Regression 4 0.00 0.00 0.00
