Effects of a Financial Transaction Tax - Do Transaction Costs Lower Volatility?: A Literature Review

(1)

Lappeenranta University of Technology School of Business

Finance

EFFECTS OF A FINANCIAL TRANSACTION TAX

– DO TRANSACTION COSTS LOWER VOLATILITY ?: A LITERATURE REVIEW 16.10.2011

Bachelor’s Thesis Tomi Sandström

d036025

(2)

1 Introduction

1.1 Backround for the research

There has been intense debate over the possible taxation of financial markets in the aftermath of the recent financial crisis. Among the suggested taxation models have been at least: a pure bank tax, a currency transaction tax, that has gained popularity as the Tobin tax and a more common financial transaction tax, later FTT, that would tax all financial transactions regardless of the currency used. Taxation of the financial markets has even been proposed in the Finnish government’s reform plan (Finnish Government, 2011). And at the time this paper is written, the European commission has made its suggestion for a transaction tax for the European Union (EC, 2011).

Currency transaction tax, later CTT, or the Tobin tax as we know it, first grabbed public attention after a lecture Nobel Laureate, James Tobin held in the university of Princeton, back in 1972. Tobin’s initial idea was that a transaction tax directed against currency transaction would reduce the profitability of speculation, hence reducing volatility on the currency market. (Tobin, 1978)

Since his proposition, many economists and researchers have grasped to the idea, but research around the topic is still contradictive. The modern renaissance of the Tobin tax is mainly due to globalisation critics and left wing politicians around the world. This is though Tobin himself has separated himself from such groups and organisations (Hodgson, 2002) and has underlined the inverse correlation between globalisation and poverty.

Since, the birth of modern derivatives markets, with their currency derivatives, it is not any more very helpful to tackle the questions around this subject by sticking strictly to the CTT. This is because the modern derivatives market has made it almost impossible to direct taxation solely on the exchange market without leaving loopholes

(4)

for traders or without affecting severe collateral consequences for other than currency traders.

Table one, which is constructed from the data gathered by Campell and Froot, (appendix) shows the list of countries, which have had some kind of transaction taxation in use in the early 1990’s. At least 16 countries from four different continents have first hand experiences of transaction taxes.

Anglo-Saxian countries such as the United States, Australia and New Zealand have had tax regimes that have all been quite moderate compared to other regimes in other countries. The Anglo-Saxian tax levels have ranged from 0,3% in Australia to 0,0033% in the United States. United kingdom has levied a fixed 2-pound burden on trades over 5000£. On top of this levy the UK tax regime contains a stamp duty tax of 0,5%. (Campell, J. & Froot, K. 1994)

Asian transaction tax regimes have typically been based on multiple compounding fees that have been avoidable when trading over the counter. For example the transaction tax system used in Singapore in the beginning of the 1990’s consisted of a 0,1% contract stamp duty, a 0,05% clearing fee and a 0,2% transfer stamp duty.

Other tax regimes in Asia have been very similar to this, with only some changes in the tax names and ratios. (Campell, J. & Froot, K. 1994)

Europe, which is commonly considered as the most prominent ground for financial taxation has also had comprehensive experience of financial transaction taxes.

European taxes have traditionally been directed against domestic trading and therefore have been avoidable when trading ex country. Tax ratios in Europe have varied from 0,05% in Italy to the 1% tax used in Finland. Sweden even tried a tax rate of 2%, but because of serious negative externalities the decided to abandon the FTT.

(Campell, J. & Froot, K. 1994)

Though a vast number of countries have experiences from FTTs, not very much research exists of their effects on volatility, revenue, trading volume and trader composition. Most of the academic literature from this field is in form of discussion

(5)

reviewed in this paper. One could generalise the debate by saying that the proponents of FTTs have a more extensive theoretical framework behind them, whereas the empiric evidence tends to support opponents of transaction taxes. As a whole, this is a field of financial economics, that is much debated, but not that much researched.

The probability of implementing financial transaction tax regulation of some sort is now greater than perhaps ever. That is why it is important that the effects as such regulation on volatility and the financial markets and the economy as a whole, is put under rigorous research, testing and analysis before legislators jump to premature conclusion on the field of regulation.

1.2 Research goals and exclusions

The main arguments in favour of a transaction tax argue, that this kind of taxation would decrease volatility in the markets and would create a way to raise funds for international aid to end poverty in the third world.

The international aid aspect will be dismissed in this debate, because the utilisation of possible FTT revenue is purely a political question and also the efficiency of international aid is under scrutiny.

Under recent years, it has become quite evident that higher transaction costs, which transaction taxes essentially are, have a negative impact on trade volume. (Umlauf, S. 1993) (Liu & Zhu, 2009) Whereas the correlation between trade volume and volatility still remain mostly unclear. William Schwert though, has noted that some researchers have found some correlation between these, but the causal relationship is unclear (Shwert, W. 1990).

Since, the most prominent argument for FTTs seems to be the view that they would lower market volatility and so increase stability in financial markets, this study intends

(6)

to shed some light to the matter related research and tries to explain the possible reasons behind the results, which still remain contradictive (OECD, 2002). By understanding the dynamics between transaction costs and volatility it can be possible to explain the reasons behind the varying results in different studies on this field.

1.3 Research methodology and data

Because of the limited scope of this article, research will be mainly conducted as a literature review. The most cited articles and studies with their writer’s views are brought up and analyzed. Also some of the criticism that has risen against the assumptions of mainstream economics is reviewed. This is mainly because these same assumptions are present in the debate over FTTs at least to some extent.

Despite of the fact that this is primarily a literature review, an empiric part is included and it will utilise quantitative analysing methods. Analysis will be conducted on a data set depicting the daily OMX-index points from the late 1980’s to late 1990’s. From these figures the daily returns and their volatilities for various time spans are being calculated. The volatilities for these time spans are being compared against the corresponding time spans before and after the Finnish FTT reform of 1992.

1.4 Structure of the research

After the introduction, in the second part of the study, the debate over FFT’s affect on volatility is reviewed and assessed trough various studies. Empirical evidence is reviewed side by side with the mainly referred arguments in the ongoing debate. A closer look at the empiric evidence is focused on tax regimes in Sweden, Japan Taiwan and China. In this part the major arguments in the ongoing debate over these implementations are also introduced.

(7)

In the third part of the study, the effects of the removal of the Finnish stamp duty in the early 1990’s are evaluated and analyzed. The overall changes in the Finnish financial markets and their regulation and the effects of the great recession of the 1990’s are also taken into account

In the fourth part results are revised against previous findings and existing theoretical framework. Conclusions are made on the basis of possible explanations of these findings. Taking in to account the urgency of the matter, it is vital to also point out important fields and topics for future studies around the matter.

(8)

2 The academic debate

2.1 The causes and features of volatility

In order to address the issue of volatility and how it is affected by transaction costs, we need to determine volatility. Volatility is essentially a measure of risk. It is the variation of the measured variable over time. Typically this variable is the return of a financial asset and its variation is measured by the standard deviation of it. Though volatility is a measure of risk, it does not make a difference between positive and negative changes in returns, it just measures the amount of change.

Volatility can be expressed mathematically by:

€

σ = ∑ ( x − x )

²

N

Where,

σ = the standard deviation

€

x

= each value in the population

€

x

= the mean of the values

N = the number of the values

Volatility can be divided to fundamental volatility and transitional noise (Hwang &

Satchell. 1997). Where fundamental volatility is the part of volatility that is caused by the arrival of new information. In principal it is caused by the change in the fundamental factors that affect any financial asset’s returns and therefore the prices.

(9)

Transitional noise is caused by the activities of irrational or speculative traders on the market, often referred as the noise traders. Noise, as Fischer Black defined it, is the irrational actions of market participants in the markets, or “uninformed trading” (Black, F. 1986). According to Black, if everybody would have the same information, nobody would trade. And because trading is what enables price discovery, it is essential that there are some differences in beliefs, which ultimately derive from the differences in information. He argued that because noise trading is based on something else than genuine knowledge, it will distort prices from their equilibrium levels. (Black, F. 1986)

The idea of equilibrium-, or correct prices, which derive from fundamental factors underlying the financial assets have since then become under criticism as shown later on in this paper. Some have also criticized the concept of genuine knowledge, in the ocean of well educated guesses, which financial markets eventually are.

It is easy to topple Black’s argument by using common reasoning: Nobody knows the future with perfect certainty, therefore nobody can have genuine knowledge about the events that happen then. And because it is the events of the future, which in essence, or at least cause most of the behaviour in financial assets returns that we perceive, it does not matter how well one knows the characteristics of a financial asset. In other words, it is not the behaviour determining characteristics of an asset that matter, but the events themselves that cause the behaviour in the first place.

Trough this view, which is contrarian to Black’s original thesis, all trading is more or less noise trading.

2.2 Volatility and Speculation

Tobin’s initial idea was to apply a CTT of 1%, which would penalize short-term currency trading, more than long-term investments. He claimed that because exchange rates are defined more by a speculative manner than trough government’s public estimates of equilibrium exchange rates, or target rates, they might give false price signals that would disturb the efficiency of capital allocation and governments macro-economical policy making (Tobin, 1978). While motives may vary between

(10)

governments and traders, it is not certain, that government’s estimates are any better than the market participant’s estimates of the so called correct exchange rates.

More over, the idea of correct or equilibrium rates defined by the underlying economic fundamentals seems to be insufficient. This is mainly because there is a considerable possibility of reflexivity between the market participant’s expectations of the exchange rates and exchange rates themselves. This is why some, George Soros among others, argue that the financial markets and the currency markets have a tendency to be naturally volatile. This is a view that is often being dismissed, even by many of the most cited academics on this field.

(Soros, 1987)

Many of Tobin’s supporters hang on to the idea that excessive speculation increases volatility. Lawrence and Victoria Summers, who are among the most cited FTT supporters, patronise this view as well as the idea of correct fundamental prices (Summers & Summers, 1989). They take Tobin’s thinking even further arguing that besides increasing volatility, short-term speculation encourages business manager to seek short-term profits instead of long-term success (Summers & Summers, 1989). If this was true, managers would knowingly have to refuse every now and then from profitable long-term projects, only on the basis of their payback time. Many find this hard to believe, considering the limited resources and therefore limited options of entrepreneurs and business managers.

Where most agree that it’s the noise traders that provide and secure the sufficient liquidity in the markets, Black’s original definition (Black, F. 1986) was broad enough to clump the uninformed traders and speculators, who should be at least somewhat informed, in to the same category. This is problematic because informed traders or professional traders can be expected to have a narrower spread tolerance and therefore less tolerance for excessive transaction costs than uninformed traders. This should be the case especially in currency markets, where informed traders trade on the basis of the prices and uninformed traders like businesses most likely trade on the basis of true need of the currency in question.

(11)

This has led Finnish economists Markku Lanne and Timo Vesala arguing that higher transaction costs penalise informed traders more than uninformed traders, thereby affecting the trader composition, which can lead to increased volatility (Lanne &

Vesala, 2006). They investigated the Deutsche Mark-Dollar and Yen-Dollar exchange rates from October 1992 to September 1993 and found out that higher transaction costs lead to higher levels of volatility. In their research Lanne and Vesala have managed to separate fundamental volatility from the data, or at least a major part of it, by taking into account the number of money-market headlines on Reuters AAMM screen on each day. This is something rarely done before. (Lanne & Vesala, 2006)

These results are in line with the findings of Charles Jones and Paul Seguin (Jones &

Seguin, 1997) and also with the findings of Steven Umlauf (Umlauf, S. 1993) as seen below. Jones and Seguin studied the effects of commission rate deregulation in the NYSE in 1975. The reform led to significant reductions in commission rates, which are essentially transaction costs. The decline in the cost of executing transactions was followed by a decline in volatility for individual shares, portfolios and for the market itself. (Jones & Seguin, 1997)

However, one must take into account that the costs of trading in NYSE were much greater back in 1975, than they are in the modern the age of online trading. Despite of this, volatility has been quite stable since the mid-19^th century in the U.S., with the exceptions of times of turmoil like the great depression and Black Monday (Schwert, 1990).

Victoria Saporta and Kamhon Kan found somewhat similar results while investigating the stamp duty regime changes in the UK. Though there were times on unusual high levels of volatility in their data, which spanned from the late 1960’s to the late 1990’s, there seemed to be no effect of the tax level change to volatility. What they found though, was that every time there was an announcement of a tax rate change, it had a significant effect on index values. (Saporta & Kan, 1997)

Steven Umlauf has studied the effects of the transaction tax regime in Sweden and his study is in many ways one of the most useful pieces of empirical evidence of

(12)

transaction taxes up to date. In the following part some light is shed on the lessons learned in Sweden.

2.3 Swedish experience of transaction taxes

Sweden introduced its transaction tax from the beginning of 1984. The tax levy was 0,5% on equity transactions and it was charged for both selling and buying, making the round-trip tax rate 1,0% (Umlauf, S. 1993). In 1986 the country doubled its tax rate on equities to 2,0%. January 1989 Sweden broadened the tax to fixed income securities, the tax levy being 0,2 basis points for a bonds with a maturity of 90 days or less, for one-year bonds 1 basis point and 3 basis points for bonds with a maturity of five years or more. Sweden also applied taxes to derivatives, but the rates were significantly lower than with equities. The maximum rate for derivatives was 15 basis points of the underlying asset’s nominal value. (Campell & Froot, 1994)

Swedish legislators planned to raise 1500 million SEK in tax revenue with the reform, but only an average 50 million SEK was raised per annum. This was mainly because a large part of the equities trade moved over seas and bond and futures were converted to forwards and debentures. (Campell & Froot, 1994)

As a result of the tax tryout 60% of the trading volume of the 11 most traded stocks in the Swedish market moved to London (Umlauf, S. 1993). During the first week of the transaction tax regime bond trading volume fell by 85%. Futures trading volumes dropped by 98% in the same time. Options and some other derivatives practically vanished from the Swedish exchange. (Campell & Froot, 1994)

The stock market reaction was also intense. The cumulative return on the Swedish All-Share index 30 days prior to the deployment of the 1% tax rate was -5,3% and on the day the tax brought to bear the index dropped further -2,2%. The day the tax was doubled the index fell by 0,8%. The relatively low figure can be explained by the fact that the markets had already long anticipated the raise. (Umlauf, S. 1993)

(13)

Because the dramatic disappointment of the outcomes of this tax regime, Sweden started lowering transaction taxes and by 1991 they had all been removed. Against these experiences it is hard to imagine how Sweden could be in favour of the EU tax regime suggested by the European Commission.

Even though the effects of the Swedish transaction taxes were severe, they did not lower volatility in the Swedish markets. On the contrary, the volatility of the trading that migrated to London fell compared to the corresponding trade with the corresponding shares in the Swedish markets. On top of this, the variance of the daily returns of the Swedish shares was greatest during the highest 2% tax regime.

(Umlauf, S. 1993) According to these findings transaction taxes are not able to lower volatility.

2.4 Experiences from Japan, Taiwan and China

The effects of transaction costs on volatility have been studied also in various markets in Asia. Shinhua Liu and Zhen Zhu studied the commission deregulation in Japan that happened in October 1999. Their findings contradict with previous empirical examinations and support the theoretical framework provided by Summer and Summers. (Liu, S. & Zhu, Z. 2009)

In their study of the Japanese commission deregulation, Liu and Zhu replicated the methods used by Jones and Seguin in 1997 and added an alternative GARCH model. Though the initial effects of the reform on transaction costs and trade volume were very similar to that in the U.S., the effect on volatility was totally different. What the researchers found, was that in the Japanese stock market lower transaction costs and higher trade volume led to significantly higher levels of volatility. (Liu, S. & Zhu, Z. 2009)

This could be explained by the change in the trader composition, as suggested by Lanne and Vesala (Lanne & Vesala, 2006). Trading in the U.S. stock market was quite different in the 1970’s than it was in Japan in late 1990’s. Partly because the

(14)

trading, it has become much easier for amateur individuals to take part in exchange trading since the 1970’s. It is possible, that ordinary individuals caused a significant part of the rise in trade volume observed in Japan. These individuals can be categorised as uninformed traders, hence explaining the rise in volatility.

Robin Chou and George Wang studied the effects of FTT decrease in the Taiwan futures exchange, TAIFEX. Their subject was the Taiwanese reduction of their transaction tax, which occurred on 1^st of May 2000. They used a structural equation framework to investigate the reform’s effects on TAIEX index futures. (Chou & Wang, 2005)

What they found was that, transaction taxes have a negative impact on trade volume and bid-ask spreads. On top of this, they found no evidence that transaction taxes can reduce price volatility, because they observed no rise in volatility after the tax reduction.

Chou and Wang argue that a rise in price volatility will change the reservation price of speculators and increase the demand for risk transfer by hedging markets participants. They conclude that this is partly responsible of the rise in trade volume that they observed. (Chou & Wang, 2005)

Nonetheless, these findings are somewhat contradictive to the study conducted in the Chinese Market. Li Zhang examined the effects of China’s increase of their securities transaction tax in May 9^th 1997 and found that the raise in the tax rate led to significant raise in return volatility and lower levels of trade volume. (Zhang, L.

2001)

On May 9^th, after the exchange had closed China announced that its was going to raise its transaction tax from 0.3% to 0.5% beginning from the following day. The announcement came without any previous warning to the investors and therefore it provides an interesting case for transaction cost related research. (Zhang, L. 2001)

However, none of these studies have managed to separate the effects of

(15)

from these studies when one has no idea of the markets related news count or atmosphere. To be precise, neither Umlauf nor Jones and Seguin had taken into account the news count related to the markets they were studying. The difference is, that Steven Umlauf was studying the same set of stocks traded in two different markets. Therefore the market and company related news affected similarly both the stocks traded in Sweden and Britain highlighting the difference in transitional volatility under the two different tax regimes.

2.5 Is the question of causality a too narrow?

Because the empirical results vary, it is justified to question the idea of a possible causal bond between transaction costs and volatility. In order for transaction costs to be the cause of volatility, transaction costs would have exist in order there to be volatility. In addition to that, transaction costs need to appear before volatility.

It is quite impossible to create a market without any transaction costs, if one uses the broad definition of the term. The fact that a market participant has to make at least some effort to engage in trading activities, can already be seen as a transaction cost.

A market with automated computer trading which does not involve humans may seem to be a market without transaction costs. However, in order for a market like this to exist, some human effort is needed to create it. So, it is actually impossible to remove all transaction costs from any market. Therefore it is impossible to out rule transaction costs as a possible cause for volatility.

This does not still mean, that transaction costs alone cause volatility, far from it.

Because the market place is as unpredictable as the human behaviour itself, the causes of volatility are likely to be at least as much psychological as mathematical.

Some see, that the way that transaction costs affect volatility depends on the market structure. Paolo Pellizzari and Frank Westerhoff argue that in a double action market like the stock market, higher transaction costs are not likely to going to dampen volatility, because a reduction in trade volume will amplify the effects of each trade on

(16)

the prices. But in a dealership market, higher transaction costs could provide stability, because it could crowd out speculative orders. (Pellizzari & Westerhoff, 2009)

In any case, it seems likely that transaction costs can affect volatility, but the relation is not necessarily this simple. Whether the effects of transaction costs on volatility depend on the market structure, as Pellizzari and Westerhoff argue, or the change in trader composition as Lanne and Vesala claim, or due to some other factor not yet recognised, it seems to be quite clear that FTTs are not the answer to lower volatilities, or at least not always, nor everywhere. More research needs to be conducted on the different contributors to volatility under different economical environments.

(17)

3 Evidence from Finland

As regard to transaction taxes and volatility, the Finnish stock market has not been investigated very thoroughly. Joakim Westerholm has conducted a research, in which he investigated transaction costs’s impact on turnover, prices and volatility during the same observation period as this study. He compared the effects of FTT removal in Finland and Sweden. (Westerholm, 2003)

Westerholm used a dynamic asset-pricing model and his findings are in line with most of the empiric evidence, that a decrease in transaction costs leads to a decrease in volatility. He also found that lower transaction costs tend to rise the level of market activity. (Westerholm, 2003)

Finland had previously a transaction tax of 1 % that was levied from exchange trading and a slightly higher, 1,6% tax rate, which was levied on over the counter trades (Finlex, 2011). The tax was removed from exchange trading from the beginning of May 1992 and was invalidated in 1996. The law that abolished the tax from exchange trading was passed in the parliament on 30^th of April, which did not leave very much time for the markets to react to the announcement. (Finlex, 2011)

When analysing the findings from Finland one must take into account, that at the time the tax was lifted, Finland was effectively in a recession.

3.1 The economic atmosphere around the tax reform

During the late 1980’s Finland enjoyed a time of strong economic growth and the economy was showing signs of overheating due to financial liberation and times of robust credit expansion. This era is commonly referred in Finland as the Casino Economy, which depicts the positive attitudes towards risk taking at the time.

The enthusiastic times of the late 1980 came to a grinding halt in the very early 1990 when the Finnish economy dived into a deep recession, caused by the simultaneous

(18)

house market and bank crises at a time when almost all of Finland’s exports to it’s biggest export country, Russia, were abolished with the demise of the Soviet union.

This kind of economic atmosphere can be difficult for analysing transaction costs effect on volatility, especially if Schwert’s findings about the historical volatility can be generalised to other markets (Schwert, W. 1990). If there is naturally significantly more volatility in times of economic turmoil and no way of separating fundamental volatility, this is a challenging era for conducting this kind of research. This is because it is the transitional part of volatility that should be affected by transaction costs, according to the reasoning of most academics.

3.2 Data and methodology

In order to investigate, the effects of the Finnish tax reform, one should have information about trade volume, daily index-value points and at least the relevant news count. However, no such data for this time period was available from the Nasdaq Helsinki Stock exchange, for this study.

Instead, data provided by the Bank of Finland was used. This data set comprises of 2510 observations of the Helsinki Stock Exchange OMX-index from 2^nd of January 1987 to 31^st December 1996. The index-values are closing values of the exchange and from them the daily revenue is calculated as the percentage of growth or decline.

The daily revenues were normally distributed.

The OMX-index contains all the stocks traded in the Helsinki Stock Exchange and they are weighted by their relative proportion in the index. Volatilities for 7 days, 15 days, one month, 30 days, one year and five years before and after the 1992 reform were calculated.

Figure one shows the development of the OMX-index with the respective daily returns over time from the observation period. The primary y-axis depicts the index

(19)

points on the left side of the chart and on the opposite side, the secondary y-axis shows the daily returns from the observation period.

Figure 1. OMX-index & daily returns

Figure two depicts the OMX-index from the same period as figure one, but instead of daily returns the secondary axis shows the seven-day volatilities from the observation period.

Figure 2. OMX-index and the seven-day volatilities

-‐8 % -‐6 % -‐4 % -‐2 % 0 % 2 % 4 % 6 % 8 %

0 500 1 000 1 500 2 000 2 500 3 000

9.1.1987 4.5.1987 20.8.1987 7.12.1987 29.3.1988 20.7.1988 4.11.1988 23.2.1989 16.6.1989 4.10.1989 25.1.1990 18.5.1990 5.9.1990 27.12.1990 22.4.1991 9.8.1991 26.11.1991 23.3.1992 16.7.1992 2.11.1992 24.2.1993 17.6.1993 5.10.1993 26.1.1994 19.5.1994 6.9.1994 23.12.1994 19.4.1995 9.8.1995 24.11.1995 19.3.1996 12.7.1996 29.10.1996

OMX Helsinki

Return

%

0 0,01 0,02 0,03 0,04 0,05 0,06 0,07

0 500 1 000 1 500 2 000 2 500 3 000

9.1.1987 4.5.1987 20.8.1987 7.12.1987 29.3.1988 20.7.1988 4.11.1988 23.2.1989 16.6.1989 4.10.1989 25.1.1990 18.5.1990 5.9.1990 27.12.1990 22.4.1991 9.8.1991 26.11.1991 23.3.1992 16.7.1992 2.11.1992 24.2.1993 17.6.1993 5.10.1993 26.1.1994 19.5.1994 6.9.1994 23.12.1994 19.4.1995 9.8.1995 24.11.1995 19.3.1996 12.7.1996 29.10.1996

OMX Helsinki

7-‐day volatility

(20)

Figure three shows the OMX-index together with the return volatility around the time of the reform giving a better picture about the possible effects of the change. The time of the reform is highlighted by a green oval on the graph.

Figure 3. OMX-index and the seven-day volatilities in 1992

From figure three we can see, that the FTT removal in Finland was followed by a moderate increase in the OMX-index, regardless of the general downward trend. This is in line with the findings of Umlauf. There is also a clear decrease in the volatility following the reform, which in line with the majority of empirical evidence in existence.

This decrease though, is followed by several months of highly volatile market conditions.

From figure 2., it is clear, that volatility tends to peak at the end of the year in the Helsinki Stock exchange, almost every year and the end of 1992 was especially volatile. Werterholm hints that the period of relatively high volatility at the end of 1992 would have more to do with the overall economic conditions than the FTT reform (Westerholm, 2003).

Table two in the other hand shows the differences in the stock exchange return volatility before and after the reform. This table summons the seven-day, 15-day, 30- day, one-month, one-year and five-year volatilities before and after the tax reform,

0 0,005 0,01 0,015 0,02 0,025 0,03 0,035

0 100 200 300 400 500 600 700 800 900 1 000

2.1.1992 14.1.1992 23.1.1992 3.2.1992 12.2.1992 21.2.1992 3.3.1992 12.3.1992 23.3.1992 1.4.1992 10.4.1992 24.4.1992 6.5.1992 15.5.1992 26.5.1992 5.6.1992 16.6.1992 26.6.1992 7.7.1992 16.7.1992 27.7.1992 5.8.1992 14.8.1992 25.8.1992 3.9.1992 14.9.1992 23.9.1992 2.10.1992 13.10.1992 22.10.1992 2.11.1992 11.11.1992 20.11.1992 1.12.1992 10.12.1992 21.12.1992

OMX Helsinki

7-‐day volatility

(21)

which was brought to bear 1^st of May, 1992. The reason why there is a 30-day volatility calculated on top of the one month volatility is that the number of trading days in each month tends to vary, mainly because of holidays. The differences in volatility before and after the reform are shown on the last row. Short-term differences are highlighted with green and the long-term differences with red.

Table 2.

Time 5 yr Vola. 1 yr Vola. 1 month vol. 30 day vola. 15 day vola. 7 day vola.

Before 0,00914374 0,009969791 0,0132071 0,011751544 0,011041333 0,007265074

After 0,012571558 0,01469998 0,007163268 0,007894061 0,007926365 0,006202871

Diff. 0,003427818 0,004730189 -0,006043832 -0,003857483 -0,003114968 -0,001062203

From the table it is clear, that the five-year period preceding the tax reform has been somewhat less volatile than the period after the reform. This is also clearly visible from figure two. However, the short-term volatilities seem to be significantly higher before the reform for all time spans from seven days to thirty-day periods. This suggests that the removal of the Finnish transaction tax did lower volatility at least in the short run. This very interesting notion is in line with most of the empiric evidence.

Like noted above, it is clearly evident from figure three, that the later part of the year 1992 was increasingly volatile. If this were because of the change in the trader composition, that could be expected to show with a slight delay, the result would contradict with the arguments made by Lanne and Vesala (Lanne & Vesala, 2006).

One must now remember that these figures represent the whole market volatility during these periods, because the fundamental and transitional volatilities have not been separated from each other. This is also the case with many other studies around this field.

(22)

4 Results and conclusions

From the volatility comparisons before and after the reform, we can see that after the abolition of the Finnish FTT, volatilities clearly decreased on the short run. These results contradict with Tobin’s original theory and what summers and summers argued, but are in line with most of the empiric evidence.

The certainty of these results still remain inconclusive mainly because there was no method available to separate fundamental and transitional volatility and the economic environment has likely had an stronger effect on volatility than the reform. However, it is fair to say, that the Finnish FTT reform of the 1992 had on decreasing effect on total volatility, when looking at the near future and past before and after the removal of the tax.

Though Westerholm’s data contained bid-ask spreads, daily highs and lows and trade volumes, even he suggests further research of the Finnish stock market for the later part of 1992. This way the economic environment’s effects on volatility would be clearer, therefore giving a better picture about the real effects of the Finnish tax reform on volatility. (Westerholm, 2003)

If transaction costs affect transitional volatility, the actual effect could be distorted in these results, because it is impossible to tell how the relation of fundamental and transitional volatilities advanced during the observation period.

It is possible, that changes in the fundamental volatility have been opposite to that in the transitional volatility, hence covering the effects of the change in transaction costs. Or, it might be that fundamental volatility’s progress has been parallel to that of the progress of the transitional volatility on the observation period, hence reinforcing the effects on total volatility. This is an issue rarely tackled in financial research around this topic, though it is one of the most remarkable single factors, which has the capability to lead to misinterpretations or simplifications of research results.

(23)

The data set used, contained no information about the trade volumes and individual trades, which could have given hints about the possible changes in the trader composition. Trade volume data goes only back to 1997, which was out of examination in this research and so late figures probably would have not given any usable information that would have been relevant to this study. Information about bid- ask spreads was available only for five years in to the past, so it could not have bee utilised either.

On the basis of this study alone, it is difficult to draw extensive conclusions about the feasibility of a FTT as a way to dampen market volatility. However, this paper is an important addition to the ongoing debate over FTTs and volatility. This study’s perhaps greatest contribution is the notion, that when addressing the effects of transaction costs on volatility, one should not investigate the changes solely in the total volatility, but instead understand that the different factors that comprise total volatility, might be affected in different ways.

If the effects of transaction costs to volatility depend on the change in the trader composition, as Lanne and Vesala suggested, then the peak in the market volatility following the removal of the Finnish FTT in the later part of 1992, should occur because a moderate rush of uninformed traders to the markets after the reform.

This seems very unlikely, if these uninformed traders mainly comprise of regular households. According to the Finland’s statistical databases, the average value of a typical household’s holdings in stocks and mutual funds almost halved from 2630 € in 1988 to 1460 € in 1994 (Statistics Finland, 2011).

The recession at the time was not the most appealing time to invest in to the stock market, at least for the risk evasive population. It appears natural that a change in trader composition resulting from a FTT reform could be different in times of severe recession compared to times of a more stable market atmosphere.

The goal of this study was to compare previous findings and bring new evidence and notable observations to the ongoing FTT debate, from a market almost never

(24)

underlying this topic as well as the field of financial economics, were put under criticism. These are aspects that can and should encourage students and academics alike to further research of the correlations between transaction costs and volatility in the Helsinki Stock Exchange. Indeed, several interesting questions still remain to be answered around this topic.

First of all, the theory of trader composition change, which was introduced by Lanne and Vesala, should be examined more carefully. A data set, with information about the market participants themselves would be very useful for this kind of research.

Also, in order to properly understand the effects of transaction costs on volatility, a comprehensive data set should be created, containing all relevant market information, statistics about the stocks and their owners, inflation and interest rates and of course all relevant legislation affecting transaction costs. Probably the best way in gather this kind of data would be with the help of some economic newspapers archives.

This kind of data set would make it possible to analyse the differences in fundamental and transitional volatilities and how transaction costs affect them. The effects could also be analysed and compared against different interest rate and inflation levels, and so possibly highlighting the different contributors of volatility under different market atmospheres.

Thankfully all the relevant data is nowadays gathered and mostly even stored. If there is to come some kind of transaction taxation in the EU or in some of the member countries, the analysing of the effects and their causes should be easier than before.

Finally, it is justifiable to underline Umlauf’s point, that it is dangerous to generalise findings of so narrow empirical evidence, when discussing policy issues of such importance (Umlauf, S. 1993) It is required that academics stay neutral on this field regardless on legislator’s decisions. While it would not cause harm either, if politicians would familiarise themselves with the existing empiria.

(25)

Empirical evidence seems to be leaning on to the contrary side against the original theoretical framework supporting FTTs. However, theories behind these results are not converging, at least yet. More testing is needed in order to better evaluate the theoretic models supporting the majority of the empiric evidence and to perhaps reveal the reasons to the still somewhat contradictive results.

(26)

List of references

• Black, F. (1986). Noice. Journal of Finance. Vol 41, Issue 3 pp. 529-543.

American Finance Association.

• Campell, J. & Froot, K. (1994) The Internationalization of Equity Markets;

Chapter: International Experiences with Securities Transaction Taxes.

University of Chicago Press. Usa.

• Chou, R. & Wang, G. (2005). Transaction tax and market quality of the Taiwan stock index futures. Journal of futures markets. [web document] [cited

10.12.2011] Available:

http://ccfr.org.cn/cicf2006/cicf2006paper/20060113210949.pdf

• European Commission (2011). Proposal for council directive on a common system of financial transaction tax and amending Directive 2008/7/EC.

2011/0261 (CNS). European Commission. Brussels. [web document] [referred 22.11.2011] Available:

http://ec.europa.eu/taxation_customs/resources/documents/taxation/other_tax es/financial_sector/com(2011)594_en.pdf

• Finnish Government (2011). Programme of the Finnish Government. Page 14.

[web document] [referred 30.10.2011] Available:

http://www.vn.fi/hallitus/hallitusohjelma/pdf332889/fi.pdf

• Finlex (2011). Leimaverolaki 662/1943. Säädökset alkuperäisinä. [web document] [cited 30.11.2011]

• Finlex (2011) Laki leimaverolain muuttamisesta 372/1992. Säädökset alkuperäisinä. [web document] [cited 30.11.2011]

(27)

• Hodgson, G. (2002). James Tobin - Embarrassed proponent of an international tax. The Guardian. London. Available:

http://www.guardian.co.uk/news/2002/mar/13/guardianobituaries.obituaries

• Hwang, S. & Satchell, S. E. (1997) Market Risk and the Concept of Fundamental Volatility: Measuring Volatility across Asset and Derivative Markets and Testing for the Impact of Derivatives Markets on Financial Markets. Accounting and Finance discussion papers 97-af37. University of Cambridge. Faculty of economics.

• Jones, C & Seguin, P. (1997). Transaction costs and price volatility: Evidence from commission deregulation. The American Economic review, vol. 87, No. 4 (Sep.,1997), pp 728-737. American economic association.

• Lanne, M. & Vesala, T. (2006). The effects of a transaction tax on exchange rate volatility. Bank of Finland research, discussion papers 11/2006.

• Liu, S. & Zhu, Z. (2009). Transaction costs and price volatility: New evidence from the Tokyo stock exchange. Journal of financial services research Vol. 36, Issue 1 pp. 65-83. Springer.

• OECD (2002). EXCHANGE MARKET VOLATILITY AND SECURITIES TRANSACTION TAXES. OECD the Economic outlook 71. [web document]

[referred 30.10.2011]. Available:

http://www.oecd.org/dataoecd/38/26/1937989.pdf

• Pellizzari, P. & Westerhoff, F. (2009). Some effects of transaction taxes under different microstructures. Working paper n. 190/2009. Department of applied mathematics, University of Venice.

• Saporta, V. & Kan, K. (1997) The effect of stamp duty on the level and volatility of UK equity prices. Bank of England working papers 1997. [web document] [cited 30.11.2011] Available:

(28)

• Schwert, W. (1990). Stock market volatility. Financial analyst journal Vol. 46, Issue 3 pp. 23-34. CFA institute publications.

• Soros, G. (1987). The alchemy of finance. John Wiley & Sons, Inc. New Jersey.

• Summers, L & Summers, V. (1989). When financial markets work too well: A cautious case for securities transaction tax. Journal of Financial services research, 3: 261-286. Kluwer academic Publishers.

• Suomen virallinen tilasto (SVT): Kotitalouksien varallisuus [web document]. Helsinki: Tilastokeskus [cited: 8.12.2011]. Available:

http://tilastokeskus.fi/til/vtutk/index.html.

• Tobin, J. (1978). A proposal for international monetary reform. Eastern Economic Journal Vol. 4, No. 3/4 pp. 153-159. Palgrave Macmillan Journals.

Available: http://dido.econ.yale.edu/P/cd/d05a/d0506.pdf

• Umlauf, S. (1993). Transaction taxes and the behavior of the Swedish stock market. Morgan Stanley International, London, England.

• Westerholm, J. (2003) The impact of transaction costs on turnover and asset prices: The case of Sweden's and Finland's security transaction tax reduction, Finnish Journal of Business Economics, vol.2:3, pp. 213-241

• Zhang, L. (2001). The impact of transaction costs on stock markets: Evidence from an emerging market. M.S. Research Paper. East Carolina University.

[web document] [cited 20.11.2011] Available: http://www.ecu.edu/cs- educ/econ/upload/zhangli.pdf

(29)

Appendix

Table 1. Transaction taxes around the world in 1991

Country Tax rate 1991 Description Notes

Australia 0,3% Transaction tax

Austria 0,15% Transfer tax Avoidable when OTC

0,06% Arrangement fee Avoidable when OTC 0,04% - 0,09% Courtage fee

Belgium 0,17% Stamp tax on buys and sells

Avoidable ex country

Finland 1% Transaction tax (round-trip) Waived, if both parties foreign.

Eliminated 1992

France 0,15% Trading tax Avoidable ex country

Germany 0,125% Boersmumsatz Steuer Residents only 0,06% Courtage tax Avoidable ex country

Hong Kong 0,25% Stamp duty

0,006% Special levy Avoidable when OTC

0,05% Exchange levy Avoidable when OTC

Italy 0,05% Stamp duty tax Avoidable ex country

Japan 0,3% Sales tax Avoidable ex country

Malaysia 0,05% Clearing fee Avoidable when OTC

0,6% Transfer stamp duty on purchases and sales

Eliminated 1992

New Zealand 0,0057% Transaction levy Avoidable when OTC Singapore 0,1% Contract stamp duty Avoidable when OTC

0,05% Clearing fee Avoidable when OTC

0,2% Transfer stamp duty Eliminated 1992

Sweden 0,5% Turnover tax Eliminated 1991

Switzerland 0,0005% Exchange fee Avoidable ex country

0,01% State tax Avoidable ex country

0,075% Stamp tax Avoidable ex country

United States 0,0033% SEC fee

United Kingdom 2 pounds Levy On trades over 5000£

0,5% Stamp duty tax Purchase only

Source: Campbell & Froot, 1994

Effects of a Financial Transaction Tax - Do Transaction Costs Lower Volatility?: A Literature Review

Table of contents

1 Introduction

2 The academic debate

€

σ = ∑ ( x − x )

N

€

x

€

x

3 Evidence from Finland

4 Results and conclusions

List of references

Appendix