Name: Thom Poorthuis Student number: 11054492 Specialization: Economics
Field: Marco and international economics Number of credit thesis: 12
Title: Effects of the 2018 steel import tariff on the stock prices of steel producing and steel
consuming firms.
Assigned supervisor: Rui Zhou Abstract
This paper discusses the effects on the stock prices of the steel tariff announced by president Trump in 2018. This will be done by researching studying the effects of the 2002 steel tariff in the literature review. To examine the effects on the stock prices for the 2018 steel tariff a data-analysis is performed. In this analysis examines if the stocks have significantly performed differently than in normal situations in the periods after the announcement of the steel tariff. This is done by finding the normal market beta for steel consuming and steel producing firms. With these beta’s a expected stock price over a month period are calculated. These expected prices have been compared to the actual prices and these differences have been tested on significance.
Statement of Originality
This document is written by Student [ Thom Poorthuis] who declares to take full responsibility for the contents of this document.
I declare that the text and work presented in this document are original and that no sources other than 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.
1 Introduction
On march the first of this year united States president Donald Trump announced an import tariff on imported steel and aluminium (CBS News,2018). This import tariff is motivated by preserving the steel producing sector. In the recent months the world is fearing for an
worldwide trade war caused by this initial steel tariff. This situation could drastically worsen if the countries affected by the steel tariff will respond with equal measures. In the May and April 2018 it seems clear that both the European union and the Chinese government (Tiezzie, 2018) will react by creating their own tariffs on products produced in the United States, this could in turn lead to even more tariffs made by the U.S government resulting in an trade war. Former trade wars as the protectionist trade war under the Smooth-Hawley act in 1930 and others have been very negative for economic growth and have proven
dramatically for all countries involved and many inhabitants, companies and traders in the countries involved are scared for the consequences (Birkley, 2018).
Although the retaliation of other actors in this possible starting trade war is an very important possible consequence for considering an import tariff this might be to broad to investigate in one study and would demand serious knowledge of law, politics and culture of the countries involved. Therefore this paper tries to focus on the direct effects of only the import tariff on steel and will not focus on retaliations of other countries who react on the tariff. The research question of this paper will be: Has there been an significant difference in the performance of steel producing company stocks and steel consuming stocks and what might be the broader economic consequences?
With all the possible negative consequences in mind it is especially interesting why the steel tariff has been created in the first place. What could be the positive effects of such an tariff and are those effects worth the gamble, it is most interesting to research all the possible effects of the steel tariff without the further consequences imposed by other
countries to figure out what could be the incentives and if they must be used as a economic tool in current and future policy. If this paper concludes with a significant negative effect of the initial tariff that might add to the discussion that it might be that trade wars are unneeded and can be better prevented by all parties involved. However if a positive effect will be found more studies can be done on the effects of import tariffs for the imposing country. An more interesting question than would be if the initial positive effect is great enough to overcome negative effects of retaliation of other countries.
The question asked will be if the stocks of both steel producing companies and steel consuming companies have under or over preformed in the a period of an month after the announcement of the implication of the steel tariff. This question will be researched in a data analysis and in the literature review the effects of import tariffs in the past will be discussed with an special interest in the effects of the steel import tariff that has been installed in 2002 by the United States. This tariff is very similar to the one today and might give good insights in the effects of the current one.
In the current upsweeping trade war it is in interest of al parties to know all the effects of a import tariff. This paper will add extra knowledge on the short term direct effect on the stock prices of an import tariff on both the industry sector that is expected to have a benefit of the tariff and the direct effects on the stock price for the industry sector that is expected to face negative effects of the import tariff. As sector with an expected positive effect the steel producing industry is chosen. For the industry with an expected negative effect the steel consuming industry is chosen.
In the first part of this paper older literature on older effects of steel tariffs and government announcements on stock prices will be discussed in a literature review. The special interest in the 2002 steel tariff is due to its great resemblance to the current steel tariff and therefore could give great insights to the effects of the current tariff.
will be performed. This analysis focusses on a period of one month after the announcement of the steel tariff. In the analysis the performance of the stocks related to the steel tariff will be researched. A comparison between their normal performance and the performance after the announcement will be done.
Preliminary result of the paper show interesting outcomes in that the steel producing industry underperformed in the period after the announcement. Also the steel consuming industry underperformed after the announcement. Although these result are not of significant value they their possible meaning is discussed in the discussion and conclusion part of the paper.
In the last part of this paper a discussion and conclusion is done. In the discussion and conclusion part the results of the data-analysis will be discussed. First the results will be compared with the expectations of this research. If the analysis has significant outcomes the intern and extern validity of these outcomes will be discussed in their usefulness for tariff policy. Also possible flaws of this type of research will be discussed that may have influenced the results. This discussion could lead to suggestions for new studies on the subject.
2.1 Effects on steel producing industry
In an interesting empirical study to the 2002 steel import tariff announced by president Bush unexpected results have been witnessed. On the day the tariff was launched there was no rise in stocks for steel producing companies, this had to do with that the tariff was already longer anticipated by the investors and they expected an even higher import tariff (Jensen, 2007) . This would assume a great deal of market efficiency on the stock markets. In most cases the stock market in total makes a better prediction of the future than do most individual investors (Mauboussin, 2002), in his model markets tend to move by the marginal best informed investor and then the bulk of investors will respond until the marginal investor reacts if a stock is over or under priced. This model however is used for normal
circumstances where the value is greatly based on future returns.
So where in the 2002 according to Jensen foresaw the import tariff and therefore stock prices even decreased after the announcement. However it was also noticed that the on the Chinese stock markets the steel producing firms lost much stock value (China alumina stocks seen hurting tags, 2001) after the first signs of the trade restrictions. It is logical that with the possibility of a reduction in profits resulting from the tariff that stock prices decrease. However it is interesting that it suggests that although investors on the American markets expected an tariff the investors on the Chinese stock markets did not expect a tariff.
In the tariff that is researched in this paper the steel tariff may also be partly expected because it have been an election promise of current president Trump (Parker, 2018). In the same article a significant negative effect on the stock prices was found on the date the WTO gave a negative judgement on the steel tariff and insisted the United States to stop the import tariff. The steel tariff of 2002 has great similarities to the 2018 steel import and an increase in stock prices before the announcement of the tariff and therefore a not significant increase in the period after the announcement or even a decrease for the steel producing company stocks on the stock market may be a possible outcome.
2.2 Steel consuming industry
As a starting point for forming ideas of the behaviour of the stocks of the steel consuming companies in the united States it is useful to understand how those companies are
influenced by the steel tariff in their business cycle and performance. An important question to whether the consuming companies are affected is if the steel producing companies will raise the steel price or will invest in increasing their production (Sedgwick, 2018). Up to 90% of the steel used in the U.S. car manufacturing sector is already coming from American steel mills. It is not the main steel that will form a problem for the car manufacturing but the special steel used for engine parts that is for most part imported from foreign steel mills. According to Sedgwick these steels are not produced in the united states and if the domestic mills do not invest to take over this production it may be still imported although to an 25% higher cost. This would mean that the production costs of the car manufacturers will increase severely and company profits may decrease. In most cases positive and negative news over future companies earnings have direct effects on the stock price (Mian & Sankaraguruswamy, 2012). Although this reaction varies strongly over time it is mostly consisted in its direction. This would conclude that if the domestic steel mills will increase price and not invest in increasing production the car manufacturing industry might be faced with an increase in production cost. This increase will lead to lower profits with in turn will lead to lower stock prices( Mian & Sankaraguruswamy, 2012). However Sedgwick concludes that the domestic steel mills will have a strong incentive to increase the production of the special steels. This would lead to a very small increase in production prices for the car manufacturing industry Sedgewick speaks of a production price increase of 0,8 % or even less for General Motors. This could still lead to an significant decrease in car manufacturing stock but severely less
than if the steel mills will raise prices and remain the same production levels. In 2002 the tariff indeed led to a rise in the steel prices for the standard car steel as for the special forms of steel (Fletcher, 2002). This was the result of domestic steel mills increasing their
production capacity but all faced with an higher minimum production cost than the foreign steel producers and this was calculated through in the end product. It is certain that in the short term the prices for normal steel and special steels will increase because the steel sector would not have enough time to increase domestic production in a short time frame. If in the long period Sedgewick is right the impact on profits may be far more smaller than expected but that is not a guarantee that investors have seen it that way. Actors on the stock market are greatly influenced by news and reputation than by actual profits (Chen & Lai, 2013) although this study has been performed on the biotech industry and therefore doesn’t necessarily is the case for the car industry it gives an good example of the influence of news and framing on the behaviour of investors and why those are far from rational.
For the steel producing companies an positive effect might be expected on first hand but in 2002 this was not the case. This was because in the time before the announcement the tariff was already expected and the investors had expected an even higher tariff and the traders’ expectations had already increased stock prices in the period before the
announcement of the actual tariff. When the news of the lower tariff reached the markets the stocks decreased and many feared the WTO ruling and the steel producing stocks
decreased even further after the ruling against the steel tariff. However in 2018 the short term consequences of the tariff on the stocks seems positive. For the steel consuming industry the stock effects are even more uncertain. Although the actual effect on costs and profits for the car industry and most steel using firms might be quite small the irrationality of investors might make the effects on the stock prices bigger than expected on basis of the rational decrease in expected earnings. This literature review is not able to give an strong direction of outcome for the data analysis of this paper.
2.3 general effects on the stock market after government announcements.
Because research done in the field of the effects on stock markets off import tariffs is quite small it is interesting to examine the general effect of government policy and especially the degree into which the market is able to forecast government policy and announcements in valuing an stock.
To investigate if investors are capable of foreseeing government policy there are various different independent cases. With the creation of the euro the interest rates of euro countries started to correlate strongly in a very small spread. The spread between bond of countries with a greater default risk and bonds of countries with less default risk. This might implicate that the investors on the stock market might have considered that there would be a common bailout in order to help a country facing default. This was foreseen by in the
Maastricht treaty and a no bailout clausula had been signed. The markets however remained sure that healthier euro countries would help and eventually the market had it right, however in the political chaos at the highest point of crisis the value of Greek bond papers sharply decreased. This case gives a mixed vision on the efficiency of the market to foresee
government policy. The market had foreseen the eventual outcome but in moment of doubt the risk averseness took the overhand (Kariya, Yamamura, Tanokura, & Wang, 2015)
The leading theoretical thinking about the efficiency of markets is the efficient market hypothesis. This hypothesis claims that the market is efficient in adjusting the value of stocks in a rational way with all the relevant information. This would implicate that the market reacts quickly to new information. This implies that a average weighted portfolio of the market will outperform a single investor always. Although this has been the case in general there are some anomalies in which the efficient market hypothesis have been disproven (Malkiel, 1989) in these cases as the 1987 Nasdaq crash and the 16th century Amsterdam tulip bubble
psychological behaviour distorted the efficient market hypothesis by speculation. In the tariff that is researched in this paper the steel tariff may also be partly expected because it have been an election promise of current president Trump (Parker, 2018). In the same article a significant negative effect on the stock prices was found on the date the WTO gave a negative judgement on the steel tariff and insisted the United States to stop the import tariff. The steel tariff of 2002 has great similarities to the 2018 steel import and an increase in stock prices before the announcement of the tariff and therefore a not significant increase in the period after the announcement or even a decrease for the steel producing company stocks on the stock market may be a possible outcome.
In an interesting paper to the relation between government announcement and stock prices it is noted that stock prices mostly react to news. In this paper positive announcement mostly leads to an increase in the stock prices of the companies affected by the news, negative announcements mostly leas to negative shocks on the exchange market for the affected firms (Hanousek & Kočenda, 2011). However this has been the leading tendons found from the data-analysis it was not always the case. In quite frequent cases positive announcements still led to negative results on the stock market. This could implicate that in some cases investors foresaw the announcements and in some cases they could not. The data analysis in this paper could thus conclude whether the investors predicted the tariff and
3 Data analysis
3.0 Methodology
In the next subpoint the data gathering and modifications will be discussed in an apart section. In this subsection the methodology of the data analysis will be discussed.
This data analysis consists of a five Step approach. In the first step the used data will be checked for stationarity. Stationarity is an underlying assumption that is needed for the use many statistical test involved in time series. This will be done with the augmented Dickey Fuller test in Stata. This test is performed on the returns of all stocks used in the later
analysis.
In the second part of this data-analysis the influence of the steel import tariff on the stocks of steel consuming firms will be analysed. This will be done by first calculating the historical beta’s of the firms stocks. Beta is a parameter used in the CAPM model. The CAPM model is a model used in finance to forecast the return of a stock based on their market sensitivity.
Ri=Rf+β( Rm-Rf)
In this model Rf stand for the risk fee return. For this constant mostly the return of US
treasury bills is used because these have very low to almost no risk. In the model Rm stands for the returns of the market proxy. In this data-analysis the proxy used will be the NYSE for the firms on the NYSE and the NASDAQ for the firms on the NASDAQ. The β is the
parameter that indicates to which extend the stock moves simultaneously with the market proxy. A beta higher than 1 indicates that a stock outperforms the market when the market proxy increases but also underperforms is the market proxy decreases. A β under 1 implies that the stock underperforms to the market proxy if the market increases but outperforms the market when the market decreases. With data of the Yahoo finance the historical beta’s for a period of five year up to the steel tariff will be calculated in a SPSS regression of the CAPM model and these beta’s will be used to calculate a expected price based on the actual return of the market in the month after the announcement of the steel tariff. This expected return is used to calculate the expected price for month after the announcement. This calculated price will be compared to the actual price on the given date and the differences will be calculated. These differences will be tested for normality with the Lilliefors test for normality. If the Lilliefors test concludes that the differences are normally distributed a T-test on the average difference can be performed to test whether the differences between the expected stock price and the actual price are of significant importance. If the Lilliefors test rejects that the differences are normally distributed than this T-test on average difference cannot be used and the Wilcoxon Signed Rank Sum test will be used to check if the differences are of significant importance.
In the third part of this data-analysis the influence of the import tariff on the stocks of the steel producing firms will be analysed. This will be done with the same steps used for the steel consuming firms.
3.1 Data gathering
All data used in this paper is downloaded from Yahoo finance. The data consist of daily datapoints over a five year period from 7-May 2012 until 9-april-2018. These spreadsheets have been downloaded into an excel spreadsheet. In the Yahoo spreadsheet several values are noted to the stocks daily information. Yahoo gives the open value, the highest value of the trading day for the specific stock, the lowest value of the specific stock, the close value of the stock and the adjusted close value of the stock. Since this data-analysis focusses on the beta’s and therefore on the stocks return only the close value are useful for the data-analysis therefore the other columns have been deleted in the excel spreadsheet. In order to see the returns of the stocks on a daily basis a new column has been constructed. In this column the new stock prices are dividend by the stock prices one day before and subtracted with 1. In this way a new column is created that shows the daily returns. In this process one datapoint is deleted because on the first day of the interval there is no past day to calculate the returns on.
3.2 Testing all data for stationarity
In the first part of this data analysis the stationarity of the returns of the stocks used will be checked because the beta’s used later in this analysis will be based on time series and an underlying assumption of many statistical tests on time series include stationarity. The test used for determine if the returns of the stocks are stationary is the augmented Dickey Fuller test. In this test the null hypothesis test for non-stationarity or better known as a random walk in the data. Although daily differences in stock often depend on the volume of the price of the stock and therefore stock differences show a nonstationary graph the relative return of a stock is almost always stationary.
The test has been performed is Stata. First the data has been downloaded in excel and the daily returns were calculated. These daily returns have been imported into Stata and tested with the ADF test. Not one of the used data sets showed signs of a random walk. All the null hypothesises have been rejected and all data sets are stationarity.
3.3 Steel consuming industry.
In the second part of this analysis there will be focused on the effects of the steel tariff on the stock market for both steel consuming and producing companies. To get an idea of how the stocks would have developed without the steel tariff an stock beta calculation has been done on nine great steel consuming companies who are registered on the NYSE. In order to make sure that this stock beta is recent enough to be of importance it is based on al historic data between the period of 7-may-2012 until 7-march-2018, the end date is chose because
President Trump announced and signed the steel tariff on march the ninth of 2018 (Horowitz, 2018). The historical stock data(Yahoo finance,2018) is organised in excel and for every specific firm its beta in accordance to the NYSE has been calculated on basis of the CAPM regression in SPSS done on the historical returns of the stocks and the market in the period of 7-may-2012 until 7-march -2018
Stock Beta based on SPSS CAPM regression of 7-May-2012 until 7-march-2018
βi
Ford Motor Company 1,144
General Motors 1,189 Boeing 1,062 Whirlpool 1,255 Anheuser 0,973 Caterpillar 1,236 DowDuPont 1,259 General Electric 1,006 Fiat Chrysler 1,215
These beta’s are used to calculate what the price of the stocks should have been according to their historic relation to the stock market. The beta’s are calculated until the date of the announcement of the tariff. On basis of the beta’s a value for the stock prices a month after the announcement is calculated to give an idea of how the stock prices would have evolved without the tariff. First the return of the market portfolio the NYSE is calculated than this value is multiplied with the specific beta’s for a stock price. The period that is used is 8 march 2018 until 9 April 2018.
On beta based expected return.
Stock Return of the market
multiplied by the specific historical beta of the firms βi×Rm
Expected price 9 April based on the return that occurs if stock follow their calculated beta precisely.
NYSE(market proxy) -0,02859629 x
Ford Motor Company -0,032714155 10,127
General Motors -0,034000988 36,553 Boeing -0,030369259 338,139 Whirlpool -0,035888343 153,082 Anheuser -0,02782419 109,260 Caterpillar -0,035345014 147,419 DowDuPont -0,036002729 68,039 General Electric -0,028767867 14,102 Fiat Chrysler -0,034744492 20,560
Column 1: Firms of stocks used
Column2: Market return Rm multiplied by specific firm beta.
Column 3: Excepted price calculated by multiplying (1+column 2)*price on 8 march
These prices are no actual current prices but the prices that would occur if the stock would have developed if it precisely continues to follow the NYSE with the calculated beta’s. These values are used to compare them with the real current stock prices of the companies. A significant lower stock price could indicate that the steel tariff had an negative effect on the stock price of these companies. In the table under this paragraph the real prices of the stocks on 9 April 2018 and the on basis of the historical beta for 9 April 2018 is summarized.
Stock actual price expected price based on calculated beta Difference Ford Motor Company 11,101 10,127 0,974 General Motors 37,830 36,553 1,277 Boeing 322,480 338,139 -15,659 Whirlpool 146,790 153,082 -6,292 Anheuser 106,018 109,260 -3,242 Caterpillar 142,363 147,419 -5,056 DowDuPont 63,689 68,039 -4,350 General Electric 12,830 14,102 -1,272 Fiat Chrysler 22,610 20,560 2,05
Column 1: Names of different firms which stocks are used. Column 2: Actual stock price on April 9 2018.
Column 3: Stock price calculated on basis of the beta and the market return Colum 4: difference between calculated price and actual price.
To verify if the stocks of these nine mayor steel consuming companies underperformed since the announcement of the tariff a statistic test is preformed to determine is the change is significant. The data is quantive and in matched pairs, this leaves two statistic tests for the central location. Both the average difference T-test and the Wilcoxon Signed test. The difference between these two tests is that the average difference T-test is done for data sets who are show a normal distribution, the Wilcoxon signed test is performed on testing the significance in the difference of the central location of nonnormally distributed data. The normality of the data set is first tested with a Lilliefors test.
In the Lilliefors test multiple steps are done to verify the normality of the data that is checked . In the first step the differences are lined up from the smallest to the biggest. In the second step the average and the standard deviation of the differences is calculated in order to calculated the Z value of all differences. In the third step the P value that underlies the Z value of every observation is calculated. In the fourth step two columns are formed the S+ and the S- column. The S- column is the percental part of the number sum of the datapoints without the data point itself. The S+ column consists of the percental sum of every datapoint including the datapoint itself. In the F(x)-S+ column the third column and the fourth are abstracted from each other and the absolute value is noted. In the F(x)-S- the fifth column is abstracted from the third. In the last to columns the test statistic of the Lilliefors test are shown on the next page.
Difference Z F(X)=P(Z<X) S+ S- F(X)-S+(X) F(X)-S-(X) -15,659 -7,0970394 0 0,111 0 0,111 0 -6,292 -1,6261520 0,0516 0,222 0,111 0,1704 0,0594 -5,056 -0,9042542 0,1841 0,333 0,222 0,1489 0,0379 -4,35 -0,4919080 0,3121 0,444 0,333 0,1319 0,0209 -3,242 0,15523009 0,5636 0,555 0,444 0,0086 0,1196 -1,272 1,30582774 0,9045 0,666 0,555 0,2385 0,3495 0,974 2,61762587 0,9956 0,777 0,666 0,2186 0,3296 1,277 2,79459596 0,9974 0,888 0,777 0,1094 0,2204 2,05 3,24607413 0,9994 1 0,888 0,0006 0,1114
Column 1: the found differences listed from smallest to biggest.
Column 2: The Z value calculated on basis of the average of the differences and the standard deviation of the differences. Column 3: The P value under the given Z value.
Column 4/5: percental summation, 4 is with the observation itself includes,5 is without the observation included. Column 6/7: 6 is the calculated P value of column 3 minus the column 4, 7 is 3 minus 5.
maximum absolute distance found in the Lilliefors analysis for D+ is 0,2385 and the maximum value of distance D- is 0,3495. The maximum distance for the level of significance(α=0,05) with 9 observations in difference is 0,271. The found value of 0,3495>0,271 is greater therefore it is on basis of the Lilliefors test confirmed that the distribution of the differences is non normal. This test result determined that the significance of the difference must be determined with a Wilcoxon Signed ranked sum Test. This test determines If the average difference found is great enough to be of significant(α=0,05) meaning. The test is shortly summarized below.
The Wilcoxon test is a test to test whether non normally distributed differences are of significant importance. The test follows a multistep approach. First the differences found will be set in absolute numbers. Than the differences will be placed from big to small. All values will than be labelled with a rank from small to big. In the last step the positive rank numbers will be summed up and the negative rank numbers will be summed up.
Difference ABS Rank Rank + Rank -
-15,659 15,659 9 9 -6,292 6,292 8 8 -5,056 5,056 7 7 -4,35 4,35 6 6 -3,242 3,242 5 5 -1,272 1,272 2 2 0,974 0,974 1 1 1,277 1,277 3 3 2,05 2,05 4 4 N=9 T+=8 T-=37
Column 1: differences from smallest to biggest. Column 2: absolute value of all differences. Column 3: given rank from small to greatest.
Column 4: all ranks summed up with positive differences Column 5: all ranks summed up with negative values.
The test statistic is Tmin(T+,T-)= 8, the critical value for the two tailed(α=0,05) is 5. In this case the test statistic is above the critical value Tmin(T+,T-)=8>5=critical value. Therefore H0 is not rejected and there is not a significant difference. This concludes that with an α=0,05 there is not enough evidence that the real stock prices on the given date and the on historic beta’s based prices differ significantly.
3.4 Steel producing industry
The next analysis is done to verify if there is a significant change in the month after the announcement of the steel tariff and the forecast on basis of historical beta’s of the stock price of eight great US steel producing companies. The companies are chosen on basis of total produced value of steel and aluminium. First the beta of the stock according to the market proxy will be determined. The market proxy’s that will be used depends on the exchange were the stock is traded. The proxy for United States steel corporation, AK steel, Allegheny Technologies, Carpenter Technology corporation, Commercial metals company and Reliance steel & aluminium will be the NYSE because the stock is traded on this exchange. The stock beta’s of Olympic steel and Steel Dynamics will be calculated with the Nasdaq as the market proxy. The beta’s are calculated on basis of dividing the covariance of the stock’s returns in relation to the returns of the used proxy and then divided by the
variance of the returns of the market proxy (Keller, 2012). The Beta’s found by using this calculation are summarized in the table below.
Stock Beta based on SPSS CAPM regression
over the historical returns from 7-may-2012 until 7-march-2018.
βi
United states steel corporation 2,038
AK steel 2,187
Allegheny Technologies 2,130
Carpenter technology corporation 1,550
Commercial metals company 1,489
Olympic steel 1,269
Reliance steel & Aluminium 1,340
Steel dynamics 1,085
These betas are on average higher than those of the steel consuming industry. This indicates that the stock prices of these companies are more volatile than the markets they are traded. As in the case of the steel consuming companies the return of the market proxy over the period form 8 march 2018 until 9 April 2018 is calculated this return will than be multiplied by the beta to find the return which would be normal to the stock on basis of the historical data and a price will be calculated. In the table below the returns of the market proxy’s by multiplying the return of the markets by the betas of the stocks a stock return is calculated.
Stock Return of the market multiplied by the specific historical beta of the firms βi×Rm
Expected price 9 April based on the return that occurs if stock follow their calculated beta precisely.
NYSE -0,0285963 X
Nasdaq -0,0642991 X
United states steel corporation -0,05828 41,704811 AK steel -0,06254 5,109179 Allegheny Technologies -0,06091 25,900152 Carpenter technology corporation -0,04432 48,560696 Commercial metals company -0,04258 23,981723 Olympic steel -0,08158 20,324621
Reliance steel & Aluminium -0,08616 83,666908
Steel dynamics -0,06976 42,604722
Column 1: Firms of stocks used
Column2: Market return Rm multiplied by specific firm beta.
Column 3: Excepted price calculated by multiplying (1+column 2)*price on 8 march
For the table above only the returns of Nasdaq and NYSE are actual data, the other
components of the table have been calculated on basis of the five year historical beta’s of the specific stocks. These calculated expected stock price have been compared to the actual stock price on 9 April 2018. The actual price has been abstracted by the expected prices and the differences are shown in the table below.
Stock Actual stock price Expected stock price Difference
(ASP-ESP) United states steel
corporation 34,740 41,705 -6,965 AK steel 4,310 5,109 -0,799 Allegheny Technologies 24,350 25,900 -,155 Carpenter technology corporation 45,969 48,561 -2,592 Commercial metals company 19,500 23,982 -4,482 Olympic steel 20,920 20,325 0,595
Reliance steel & Aluminium
83,540 83,667 -0,127
Steel dynamics 43,680 42,605 1,075
Column 1: Names of different firms which stocks are used. Column 2: Actual stock price on April 9 2018.
Column 3: Stock price calculated on basis of the beta and the market return Colum 4: difference between calculated price and actual price.
As in the case for the steel consuming company stocks a Lilliefors test for normality must be performed before a test can be done to determine if the difference can be seen as significant. The Lilliefors test is shown in the table below.
Difference F(X)=P(Z<X) S+ S- F(X)-S+(X) F(X)-S-(X) -6,965 0,000 0,125 0,000 0,125 0 -4,482 0,001 0,250 0,125 0,249 0,124 -2,592 0,164 0,375 0,250 0,211 0,086 -0,799 0,829 0,500 0,375 0,329 0,454 -0,155 0,951 0,625 0,500 0,326 0,451 -0,127 0,954 0,750 0,625 0,204 0,329 0,595 0,993 0,875 0,750 0,118 0,243 1,075 0,999 1 0,875 0,001 0,124
Column 1: the found differences listed from smallest to biggest.
Column 2: The Z value calculated on basis of the average of the differences and the standard deviation of the differences. Column 3: The P value under the given Z value.
Column 4/5: percental summation, 4 is with the observation itself includes,5 is without the observation included. Column 6/7: 6 is the calculated P value of column 3 minus the column 4, 7 is 3 minus 5.
For the test the biggest value determines the normality. The biggest value of 0,454 exceeds the critical value 0,454>0,285 therefore the differences of the expected and actual stock prices are of the steel prices are significantly(α=0,05) not normal distributed. Because of this non normal distribution a Wilcoxon signed rank test is performed. This test is shown in the table below.
Difference ABS difference Rank
-6,965 6,965 8 8 -4,482 4,482 7 7 -2,592 2,592 6 6 -0,799 0,799 4 4 -0,155 0,155 2 2 -0,127 0,127 1 1 0,595 0,595 3 3 1,075 1,075 5 5 N=8 T+=8 T-=28
Column 1: differences from smallest to biggest. Column 2: absolute value of all differences. Column 3: given rank from small to greatest.
Column 4: all ranks summed up with positive differences Column 5: all ranks summed up with negative values.
The test score of the test is Tmin(T+,T-)=8, the critical value is 3. In the Wilcoxon signed rank sum test a difference is significant if the test score is under the critical value. This concludes that there is not strong enough evidence(α=0,05) to conclude that the difference between the expected stock prices and the actual stock prices is significant.
4 Discussion
In the data analysis un significant results have been found. This however does not mean that there has not been an difference in performance for both the steel producing sector and the steel consuming sector.
The steel producing sector showed quite big decreases in stock price. This looks in first hand to be in contrast with the study on stock prices and government news (Hanousek & Kočenda, 2011). This because the import was suggested to be good news for the steel producing sector and in the study was concluded that news affects stock prices mostly in logical ways.
On the other hand it shows great similarities to the 2002 steel tariff. According to Jensen the stock investors than already foresaw the steel tariff and overvalued the steel producing stocks for that reason in the months before the announcement. This seems a very likely explanation for the current underperformance of the steel producing stocks. This would imply that investors where able to forecast the import tariff on basis of older announcements and promises of president Trump. This would very well be possible because helping the steel industry was one of the major campaign arguments of president Trump.
Although Jensen’s paper gives an good explanation for the found results of the data analysis for the steel producing sector it forms a difficult question for the steel consuming industry. The steel consuming firms showed a decrease on average in the month after the announcement. If the explanation that Jensen’s paper gives for the effects on the stock market is correct and indeed the investors foresaw the tariff why would they not undervalue the steel consuming firm stocks in the months before the announcement. It would not resemble rational stockholders behaviour if the investors foresaw the tariff but only acted in advance for the stocks of steel producing firms.
A possible explanation for the decrease in the steel consuming industry could be that in the days after the announcement many spokespersons for the steel consuming industry claimed in the media that by the increase in costs the profits will go down substantiality. This news of decreasing profits mostly has a very strong negative effect on the stock prices of the firms (Mian & Sankaraguruswamy, 2012). So it could be the case that although investors already foresaw the import tariff that they where simply extra influenced by the news reporting over the severe effects for the steel consuming industry.
Another explanation for the anomaly the steel consuming stock decrease makes can be found in the research of the efficient market hypothesis. In the study of Malkiel he concluded that although the market is more often than most experts capable to take all possible information in making the good valuation decisions it is also sometimes terrible wrong in valuing a stock. It could be the case that although stock investors foresaw a tariff they were not possible to determine how big this tariff would be and it possible
5.0 Conclusion
The results of the Wilcoxon test for difference of the stocks for steel consumers was not of significant value. Although the stocks on average over the period have underperformed this difference was not great enough to be of significant value. It is important to note that the expected values that have been used as comparison material have been based on the historical relationship between these stocks and the exchange market. Although this relationship is based on five year of daily stock returns it is not always stable. The beta shows some short term differences that may have influenced the expected price. Also must be stated that the expected price was only based on the historical beta. Firm specific news and managerial decisions could have a great influence on the stock price of these specific firms. Therefore it would be very useful to do a broader study on news items concerning these firms and if those news items have had an influence on the stock price of these firms. This could in some cases make that some firms are not useful for the data-analysis. Also of great importance for the stock prices of these firms have been the widespread media
attention the tariff received. In the month after the announcement there have been very much uncertainty about the implication of the tariff. Manny countries tried to get there selves an exemption. In the period this data-analysis researched there have been many negotiations between the United States and the European Union and China. These negotiation and the media attention they drew might have had a strong impact on the stock prices of the companies concerned. In order to examine the effects on the stock price of the tariff it is important to examine the effects of these media items on the steel consuming industry. It is also of great importance to reflect on the point that stock prices show great volatility in short term. It is not impossible that if another date was chosen for the ending date of the analysis that other results would have been achieved. Therefore it would be useful to do a longer period of time series analysis in order to examine if the stocks underperformed in their normal relationship to the market over a longer period of time.
The results of the data-analysis of the stocks of the steel producing industry was also not of significant meaning. However not significant there have been differences between the expected stock prices and the actual stock prices. The announcement of the steel tariff was ought to be positive news for the steel producing sector. The differences found between het expected price based on the historical beta and the actual price were on average negative. This is in contrast with what was expected however it makes a strong case for market efficiency. It is possible that investors in the stock market had already expected government action. As the protection of the steel industry has been a key point in president Trump electoral campaign the public might have expected a tariff. This expectations could have translated in higher stock prices in the period after the election of president Trump. On the day of announcement investors could have been disappointed with the actual government measures because the investors had expected even higher import tariffs or other measures of government protection of the steel industry. For further research it would be useful to determine if the stocks of the steel producing sector have outperformed in comparison to their normal market relation. If this would be the case that would make an strong evidence that markets are often good in predicting government actions to a certain degree and
therefore supporting the efficient market hypothesis. Another point that might be important to be studied more intensively it the effect of news on the stock prices of the steel producing industry. In the period used for the data-analysis many uncertainty arose over the precise implication of the tariff. This might have the same implications as the stock prices of the steel consuming industry and result in lower stock prices caused by much uncertainty. The effects of news on the negotiations between the United States and other trading partners and their effect on the stock prices of steel producing firms must be further researched.
The differences have not been big enough to support significant effects. Because of this results this paper cannot be used in order to advise policy. However this study has made the start of some interesting insights in investor behaviour that could be the beginning of
interesting further research. The effects on the stocks might indicate that the investors foresaw the government actions. In a bigger and longer term research that focusses not only on a one month period but on longer and other relevant points in time this could be
researched. In a study that focuses on the performance of the stocks in the period after the election of president Trump until the announcement of the tariff it could be determined if the investors foresaw the tariff.
Another short coming of this research was the short time after the
announcement that it has been written. In the period of time this paper was written the news around the tariff was still very turbulent. Uncertainty about the precise implication of the tariff caused big daily shocks on the stock market concerning both the steel producing and steel consuming industry. It would have been more useful to do this data analysis at the moment that there is more clarity about the actions of all parties involved.
In short a new study on the tariff should be performed in order to investigate if investors foresaw the tariff. This study must be done in a later stadium and should focus on more on the effects of media on the stock prices and on the possibility that investors foresaw the tariff. This study could be an extra confirmation of the efficient market hypothesis or it could lead to more interesting studies if it would reveal an special anomaly.
References
Birkley, J. (2018, March 2). Sorry, President Trump, Trade Wars Are Bad And Impossible To Win. Retrieved from Forbes: https://www.forbes.com/sites/johnbrinkley/2018/03/02/trade-wars-are-bad-and-impossible-to-win/#1423aa7e1add
CBS News. (2018, Maart 1). Trump announces steel and aluminum tariffs. Retrieved from CBS news: https://www.cbsnews.com/news/trump-institutes-steel-and-aluminum-tariffs-live-stream/ Chen, T., & Lai, M. (2013). Are Investors Rational? Evidence on the Impact of Industrial Framing
Reclassification on Stock Market Reaction. Journal Of Behavioral Finance, 1-8. China alumina stocks seen hurting tags. ( 2001). American Metal Market, 4.
European Central Bank . (2018). Transmission mechanism of monetary policy. Retrieved from ECB: https://www.ecb.europa.eu/mopo/intro/transmission/html/index.en.html
Fletcher, S. (2002). New steel tariff could trigger like action on OCTG. The Oil and Gas Journa, 37. Foucault, T., & Fresard, L. (2013). Learning from peers' stock prices and corporate investment.
Journal of Financial Economics, 554-577.
Hanousek, J., & Kočenda, E. (2011). Foreign News and Spillovers in Emerging European Stock Markets. Review of International Economics, 170-188.
Hau, H. ,. (2013). Real effects of stock underpricing. Journal Of Financial Economics, 392-408. Horatiu, D. (2013). THE ASSET PRICE CHANNEL AND ITS ROLE IN MONETARY POLICY TRANSMISSION.
Annals of the University of Oradea: Economic Scienc, 445-454. Horowitz, J. (2018, march 9 ). money.cnn.com. Retrieved from cnn:
http://money.cnn.com/2018/03/08/news/companies/companies-steel-aluminum-tariffs-list/index.html
James Dow, G. G. (1995). Stock market efficiency and economic efficiency : is there a connection? Cambridge, Mass. : National bureau of economic research.
Jensen, N. (2007). International institutions and market expectations: Stock price responses to the WTO ruling on the 2002 U.S. steel tariffs. The Review of International Organizations, 261-280. Kariya, T., Yamamura, Y., Tanokura, Y., & Wang, Z. (2015). redit Risk Analysis on Euro Government
Bonds-Term Structures of Default Probabilities. Asia-Pacific Financial Markets, 397-427. Keller, G. (2012). Manegerial Statistics. South-Western CENGAGE Learning.
Kosfeld, R. (2002). Asset price channel and financial markets . Journal of Economics and Statistics, 440-462.
Malkiel, B. G. (1989). Is the Stock Market Efficient? Science, 1313-1318.
Mauboussin, M. J. (2002). REVISITING MARKET EFFICIENCY: THE STOCK MARKET AS A COMPLEX ADAPTIVE SYSTEM. Journal of Applied Corporate Finance, 47-57.
Mian, G. M., & Sankaraguruswamy, S. (2012). Investor sentiment and stock market response to earnings news. Accounting Review, 1357-1364.
Parker, B. (2018, Maart 5). Seatrade maritime news . Retrieved from
http://www.seatrade- maritime.com/news/americas/trump-steel-tariffs-reinforcing-bars-campaign-promises-and-thinking-out-loud.html
Qi chen, I. G. (2007). Price Informativeness and Investment Sensitivity to Stock Price. The Review of Financial Studies, 619-650.
Romer, D. (2012). Advanced macroeconomics. New York: McGraw-Hill.
Sedgwick, D. (2018). U.S. tariff on steel targets hot segments; Will U.S. producersraise prices or invest? Automotive News, 1-22.
Tiezzie, S. (2018, June 5). Brace Yourselves: The US-China Trade War Is About to Begin. Retrieved from the diplomat : https://thediplomat.com/2018/06/brace-yourselves-the-us-china-trade-war-is-about-to-begin/
WANG, P. B. (2011). A Unified Theory of Tobin's q, Corporate Investment, Financing, and Risk Management. Journal of finance, 1545-1578.
YahooFinance . (n.d.). Retrieved from Yahoofinance.com :
https://finance.yahoo.com/quote/FCAU/history?period1=1336341600&period2=152547120 0&interval=1d&filter=history&frequency=1d
