Inflation and Stock Returns in Nigeria

largeness and billet Returns in NigeriaThis ask experiment wholey examines the consanguinity mingled with swelling and demarcation tax breaks in Nigeria during 1997-2006. The get wind revolve ab fall outes on different econometric standards to investigation this family victimisation monthly info of the All administer Price Index from the Nigerian gestate supercede and Nigerian Consumers Index. The undecomposable OLS reverse firmness adumbrates that the residuals ar stationary, which implies that rip productions and pretension argon co integrated. on that pointfore we brush aside conclude that there is a desire strike human birth among inventory returns (LOGASI) and puffiness (LOGCPI).The Engel co- integrating egresss reveals that there is long lay out kin mingled with pompousness and linage returns .the sphere further goes on to the determine the causal long slope family blood using the wrongful conduct subject area Model (ECM). This a rticle offers evidence of a compulsory descent amidst inventory shell out returns and fanf atomic number 18. This expiry confirms that declivity returns act as a hedge against ostentatiousness.CHAPTER ONEINTRODUCTION1.1 Background to the StudyThe orgasm of oil boom in Nigeria in the early 1970s, has led to the dissymmetry of line of reasoning prices. This has been attributed to many factors such as budget deficit monetization, inflow of foreign capital from crude oil sales and financial market places mental institution of lavishness private national credit.Since early 1970s, swelling rates in Nigeria has been highly unstable the high fanf arary change was in excess of 30 percentage. This is evident in the high correlation of money tote up ripening and high flash due to the fact that authoritative number stintingal increment is less in authoritative term to money growth. This can be observed from the growth in money supply and around structural factors such as supply shock absorbers arising from famine, unfavorable terms of divvy up and devaluation of currency. Furthermore, Structural Adjustment Program (SAP) introduced by the goernment in the slowly 1980s to a fault accounted for the step-up in price level in the deliverance. Consequently, splashiness in Nigeria has everyplace prison term responded to structural changes. These changes can be characterized into 4 expirations found on the pattern and events that occur at that close.The first period of term of fanfareary outgrowth in Nigeria was detect from 1974 to 1976 lump increased by 30 percent. This puffinessary pressure was as a result of the pastimeHigh cost of agricultural produce ca utilize by drought in the Northern part of Nigeria,Excessive oil revenue monetization,increase in wage rate based on the recommendation of the Udoji commission of 1974, Folawewo (2005), and constitution-making in stabilityThe second period was from 1983 to 1985 when ostenta tion rate reached 40 percent. This period noniced very little economic growth, The Nigerian government was low intense pressure from debtor groups to accept International Monetary coronation trust conditionalitys of devaluation of internal currency beca role government debt has increased to a higher place 70 percent sequence excess money growth was around 41and 43 percent. This period excessively witnessed poor external trade per do workance.CBN (, 2006)The third period was from 1987 to 1989 when rising prices rate hovered around 35 percent. During this period, the thriftiness experienced high inflationary pressure brought about by fiscal expansion noniced in the 1988 budget, the debt for equity swaps con fluctuation method adopted by the Government of Nigeria and the drastic contraction in monetary policy, all accounted for this change that span through to the early 1990s.Finally, the fourth period occurred among 1993 and 2000, as a result of fiscal deficit expansion whi ch caused a 70 percent increase in money supply with a knock-on achievement on domestic credit of the private sector of the economy.CBN, (2006)Overall, inflationary pressure can be largely attributed to structural factors such as literal income reduction caused by mutant in oil revenue, high titular wages and debt stipulation in form of expansionary fiscal deficit. These invariably mean that over the years, variance in commodity price is a normal feature of the Nigerian economy.One choose commodity considered in this study is the capital market farm animal, i.e. the Stock market. Stocks listed in Nigeria are traded on the floor of the Nigerian Stock Exchange (NSE) age the Securities and Exchange Commission (SEC) is the apex regulatory physical structure which oversees the activities and affairs of the major players on the floor of the Stock Exchange.The Nigeria Stock Exchange was established in September 15, 1960 but commenced business on June 5, 1961 with 19 securitie s listed and traded on the Lagos Stock Exchange. Based on the recommendation of the Government pecuniary System Review Committee in 1976, the Lagos Stock Exchange was renamed and make part of the Nigerian Stock Exchange in December 5, 1977. The Nigerian Stock Exchange has nine branches established in major commercial-grade cities in Nigeria. The main tack of strains of large enterprises are traded in the Nigerian Stock Exchange while small and medium scale enterprises are listed and traded in the Second tier Securities Market (SSM). From 1963 to 1990, the Nigerian declivity stand in witnessed an overwhelming increase in government air which exceeded the equities of industrial companies s gutter this trend changed from 1991. The apprize of equities of industrial companies increased to billions of Naira, while government have a bun in the oven traded on the Nigerian Stock Exchange was worth millions of Naira this decrease continues till date, a development to the deregulatio n of the economy.Despite the increase in market capitalization noticed in the economy at that period, the ratio of this summate to the Gross Domestic Product and Gross Fixed Capital composition was still low. This increase was amidst 4.8% and 25.4% for gross domestic product while the ratio for gross fixed capital formation is between 28% and 55% from 1963 to 1990 (CBN, 2006). The ratio of market capitalization in the gross domestic product and gross fixed capital formation increased geometrically from 1990 to 1995. Although there was decrease in the share of market capitalization in gross domestic product and gross fixed capital formation, the return on investment did not follow the identical pattern. This decrease noticed at that period was caused by a banking crisis in which a fall of 26 banks were liquidated in 1998. However, with the recapitalization of the banking sector in 2005, the industry remains the closely active participant in Nigerian contrast market up till dat e. The trend in Nigeria Stock Exchange causes the price and return on furrows to be highly volatile.1.2 Problem StatementPrice stability is essential in find out whether an economy is stable or not. Inflation which is the constant increase in price nominates question in the economy uncertainty makes both domestic and foreign investors un ordaining to invest. In Nigeria inflation has led to increase in nominal engagement rates which come upon the value of interest payment of banks and financial institutions. Furthermore, determination of the problem caused by inflation depends upon the degree in which inflation is anticipated correctly or not. If inflation is anticipated correctly and the monetary authority is seen to be credible, the fluctuation in price would be managed publicationively but if inflation is unanticipated, some(prenominal) economic agents allow gain while new(prenominal)s will lose. Unanticipated inflation impact opposely on saving ability of the citizens and as a result, low saving leads to a fall in the demand for stocks and equities as financial wealth. This decrease in demand causes the price of equities to fall thereby reducing returns on equities and stocks.Furthermore, the prices of stock determine how effective and efficient the stock market allocates shares and equities based on preference and availability of market information. subjoin or decrease in price of stock create uncertainty for the investors and in turn affect the demand and supply of stocks. Therefore, general increase in price level may affect peoples potential investors investment decision which has electro disconfirming impact on the total returns on stocks in the economy at large. This situation is prevalent in the Nigerian economy therefore there is the impoverishment to examine the effect of inflation on stock returns and its implication on investment. The fisher cats guess (Fishers effect) suggests that stocks or equities hedge or evade inflation, empi rical investigation suggest that inflation and stock returns are negatively related. This study will be looking at family race between inflation and stocks in Nigeria.The study of this kind is essential in improving and in the understanding of stock markets, thus providing standards for decision-making about addition allocation.This study contributes to the existing literary works by providing evidence for whether inflation affects stock returns both in the long run and in the nobble run.1.3 excuse for the StudyDespite the large number of empirical studies on the relationship between inflation and stock returns, there is no general consensus on the causal direction of this relationship. Empirical works as Nelson (1976), Shwarts (1977), Fama (1981), Geske and rolling wave (1983), Gultekin (1983), marshal (1992), Bakshi Chen, (1996), Zhao (1999), Chatrath et al (1997), Spyrou (2001), Omran and Pointon (2001), Crosby (2001), Gal disposeher and Taylor (2002) and Floros (2002), su ggested a negative relationship between inflation and stocks while Boudoukh and Richardson (1993), Graham (1996) and Choudlery (2001) in different studies take the opposing view, i.e. that there exists validatory relationship between inflation and stock returns.However, most of these studies were carried out in industrial nations and some selected developing countries most especially Latin American countries. Specific studies on the exact relationship between inflation and stock returns in Nigeria have not been explored rigorously. Furthermore, considering the negative impact of inflation on prices of commodities in Nigeria coupled with the volatility of stock returns, this study seek to abide a rigorous analytic thinking of the propellings of inflation and its implication on stock returns in Nigeria using an wrongful conduct Correction Model to create a parsimonious and encompassing model that would luff both short-run and long-run relationship between inflation and stock r eturns in Nigeria.1.4 jut of StudyFollowing the introductory remarks in chapter one, chapter devil will brush up the existing literature on this subject. While chapter three will focus on the theoretical framework, methodology, model specification, estimation technique and sources of info. The summary of result of the empirical analysis is presented in chapter four while the study will be rounded up in chapter five with summary of findings, policy implication and conclusion.CHAPTER TWOLITERATURE REVIEW2.1 IntroductionSection 2.2 of this chapter discusses the underpinning theories of inflation and stock returns. Section 2.3 examines the empirical literature review on inflation and stock returns this is to help identify the link between inflation and stock returns. Finally section 2.4 examines the methodological literature on inflation and stock returns.2.2 Theoretical Literature Review on Inflation and Stock ReturnsThe Fisher system suggests that there is a positive relationship between interest rates and inflation. (Berument Jelassi, 2002) Fisher (1930) argues that nominal interest rate is entirely a scrape of the existing information in relation to the kindredly afterlife values of the rate of inflation. This hypothesis has come to be known as the Fisher effect in the economic literature it states that evaluate nominal rates of interest on financial assets should move one-to-one with expected inflation. Choudhry (2001) Fisher hypothesis, in its strict sense, predicts a positive homogeneous relationship of degree one between stock return and inflation. (Luintel Paudyal, 2008)The proxy-hypothesis was introduced by Fama (1981) to pardon the predominance of negative stock return-inflation trend. The main principle on which Famas version of the proxy-effect hypothesis is based on is the observed negative relationship between inflation and stock returns which step forwards to be spurious since this relationship is a result of the positive relationship that exist between stock returns and expected economic action at law and an inverse relationship between expected economic activity and inflation. Inflation simply serves as a proxy for expected economic activity in a statistical relationship between stock returns and inflation. (Lee U. , Monday, June 22 1998)The proxy hypothesis states that the negative relationship between inflation and stock returns is spurious and really merely proxies for the positive relationship between stock returns and real changeables. Previous renderes of the proxy hypothesis have used actual values sooner of forecasted values for the real activity variable. (McCarthy, Najand, Seifert, 1990) did not find a support for the proxy hypothesis using solely forecasted variables.Gonedes (1981) the harm to use indexation means that real income tax rates will vary directly with rates of inflation. This substantive effect of mere bookkeeping methods is frequently predicted even though it is known to have som e ominous implications. This is the tax effects of inflation hypothesis.2.3 Empirical Literature Review on Inflation and Stock ReturnsThe empirical literature on the impact of inflation on stock returns records major contribution by different scholars over the years. But the empirical evidence provided by most of these studies has been mixed, and a consensus has not yet emerged. While studies like Pierrel and Kwok (1992), Geske and Roll (1983), Floros (2002), Ugur (2005), Yeh and Chi (2009), Pesaran et al (2001), hideaway Haan (2000), Crosby (2001), Syros (2001), Roohi and Khalid (2002) among early(a)s have found a negative relationship between inflation and stock returns Boudoukh and Richardson (1993), Graham (1996), Choudhry (2001), Patra and Posshakwale (2006) and Lee et al (2000) among others reported positive relationship between these variables.Concerning the review of the approaches of modeling the effect of inflation on stock returns, Pierrel and Kwoks (1992) estimates an d try outs the alternative versions of hypothesis that explain the relationship between these deuce variables. The study calls distributed lags in differentiate to empirically arrive at a dynamic structure of inflation. Pierrel and Kwoks concluded that this dynamic structure conform to Fama (1981), Benderly and Zwick (1985), and Geske et al (1983) hypothesis that suggest a negative relationship between inflation and return on stocks.Yeh and Chi (2009) time-tested the validity of the various Hypotheses that explain this relationship. The empirical result of this study on 12 OECD countries raises that these countries exhibit a short-run negatively hearty co-movement between stock returns and inflation. Moreover, countries like Australia, France, Ireland and Netherland do not display a long-run relationship between the two variables in equilibrium. This result is consistent with the hypotheses of Fama (1981), Modigliani et al (1979) and Feldstein (1980) which suggested that an in crease in inflation reduces real returns on stock. This result is also in line with Caporale and Jung (1997) and Rapach (2002). They argue separately that there exist a negative probatory effect of inflation on real stock returns after controlling for output shock and that inflationary trends do not erode returns on stocks.The Fishers Hypothesis was tested by Spyros (2002). His results reflect a contrary view that returns on stocks hedges inflation. This study establishs that there is negative but not statistically significant relationship between inflation and stock returns in Greece from 1990 to 2000. In this same vein, Floros (2002) carried the same study on Greece economy and concluded that inflation and stocks in Greece should be treated as independent variables because the result of the various test conducted show that there is no relationship between inflation and stock returns in Greece. Crosby (2001) check up ons the relationship between inflation and stock returns in A ustralia from 1875 to 1996 and found out that the Australian economy does not experience permanent changes in inflation or stock returns. The result shows that there exist short-run negative relationships between these two variables that depend on the period of time that is considered.On the contrary, Lee et al (2000) examine the impact of German hyperinflation in the twenties on stock returns. This result of this study show that the hyperinflation in Germany in early 1920s cointegrates with stock returns. The fundamental relationship between stocks returns and both realized and expected inflation is highly positive. They concluded that common stocks appear to be a hedge against inflation during this period. Choudhry (2001) in his study on the impact of inflation on stock returns in some selected Latin and commutation American countries (Argentina, Chile, Mexico and Venezuela) from 1981-1996, reveal that there is one- to-one relationship between the accredited rate of nominal retur n and inflation for Argentina and Chile. Their result also reveals that the lag values of inflation affect stock returns and this result generalize that stocks act as a hedge against inflation.Patra and poshakwale (2006) conducted a study on the impact of economic variables on market returns in Greece from 1990 to 1999. Empirical results show that some macroeconomic variable like money supply, inflation, volume of trade and exchange have both short-run and long-run relationship with stock price in equilibrium in Greece while there was no short-run or long run relationship noticed between exchange rate and stock prices.Ugur (2005) in a study on the effect of inflation on return on stocks in misfire from 1986 to 2000 reveal that expected inflation and real returns are not correlated. The results suggest there is a negative relationship between inflation and stock returns which may be caused by the negative impact of unexpected inflation on stock returns. This results did not contradi ct Fisherian hypothesis because of the non correlation of inflation and real returns but the results is in line with the proxy hypothesis since a negative significant relationship exist between the two variables. Aperigis and Eleftheriou (2002) results also concurred that there is negative link between inflation and stock returns in Greece than in interest rate and stock returns. Similar study like Adrangi et al (1999) and sellin (2001) also support the proxy hypothesis. Khil and Lee (2000) in their study on ten pacific-rim countries and the US that all the countries except Malaysia reveal negative relationship between inflation and stock returns.The tax-effects Hypothesis which asserts that there is negative relationship between inflation and stock returns was tested by Geske and Roll (1983). Empirical result from the reveal that random negative or positive real shock affects stock returns which in turn, signal higher(prenominal) or lower unemployment and lower or higher corporate earnings. This has effect on the personal and corporate tax revenue leading to increase or decrease in the treasury through borrowing from the public. The economy compensable for this debt by expanding or contracting money growth and this would lead to higher or lower inflation. They concluded that random shocks on stock returns are both fiscal and monetary in nature in the U.S.A.Roohi and Khalid (2002) considered the efficacious Market Hypothesis and Rational Expectation Theory to investigate the effect of inflation on stock returns. Empirical results of the study suggest that the relationship between real stock returns, unexpected inflation and unexpected growth are negatively significant. They concluded that the control of real output growth makes the negative relationship between these two variables to disappear over time.2.4 methodological Literature Review on Inflation and Stocks ReturnsThe empirical relation between inflation and stock returns has been investigated through various approaches since the 1970s. Spyros (2001), adopted Vector-Auto regressive (VAR) model and the co integrating test to confirm if there was any relationship between inflation and stock returns in Greece. Pierrel and Kwok (1992) investigated the relationship between stock returns and inflation in the United State between 1962-1992 using Vector- Autoregressive (VAREC) model, and Granger cause, Crosby (2001), used Vector-Autoregressive (VAR) model, modal(a) Least Square (OLS) and correlation analysis to examine the relationship between inflation and stock returns in Australia from 1875-1996.Floros (2002), investigated the relationship between stock returns and inflation in Greece from 1988-2002 by considering both the lag and lead periods of inflation and stock returns using Ordinary Least Square (OLS), Johansen Cointegration Test and Pairwise Granger Causality Test. In this same vein, Ugur (2005) used the Ordinary Least Square (OLS) and standardised Granger Causality to examin e the relationship between inflation, stock returns and real activity in Turkey.Choudhry (2001), estimate the impact of inflation on stock returns in some selected Latin and Central American countries using the Auto-Regressive interconnected Moving Average (ARIMA), unit lineage test and spectral reasoning backward model. Lee et al (2000) and Geske and Roll (1983), also used ARIMA, OLS and unit root test to investigate the effect of German hyperinflation and stock returns, and the impact of inflation on stocks returns in the USA respectively.Patra and Poshakwale (2006) on the other hand, used the error Correction Model (ECM), Johansen Cointegration Test and Pairwise Granger Causality Test to show if economic variables such as money supply interest rate, exchange rate, volume of trade and stock prices have impact on stock returns.Yeh and Chi (2009) in their study on 12 OECD countries measures correlation at different forecast horizon by using Autoregressive Distributed Lag (ARDL) shrink test, unit root test and confidence interval method to investigate the inflation illusion hypothesis that suggest that there is negative relationship between inflation and stock returns. Pesaran et al (2001) and Den Haan (2000) also employ the same technique and arrive at the same result.This study examines the relationship between inflation and stock returns in Nigeria. Furthermore a test is carried out to see if theres a cointegration and causality within these variables. Methods used in this study are explained in chapter three. This study fundamentally aims to analyses the above relationship for a period of 1st of January 1997-31st of December 2006 .monthly values of the Nigerian Stock Exchange (NSE) and Nigerian Consumers Price Index (CPI). CPI was collected from the Central Bank of Nigerian statistical bulletin (2006), while (ASI) All grapple Index was collected from Nigerian Stock Exchange data bank.The reviews of literature above reveal that there are basically four major hypotheses discussing the relationship between inflation and stock returns. These theories are Fisherian hypothesis, proxy hypothesis, tax-effect hypothesis and inflation illusion hypothesis. Considering the level of price stability in Nigeria over the period of our study, the study seeks to adopt Fisherian hypothesis which suggest that stock hedges inflation. This is based on the fact that literature suggests that the price of stock is a major determinant of stock returns which is affected positively by expected or unexpected inflation (consumer price index).CHAPTER THREEMODEL SPECIFICATION AND methodological analysis3.1 IntroductionThis chapter covers the theoretical framework, specification of the models utilized in the study as well as the methodologies that will be adopted. Accordingly, the estimation procedures, and data requirements types and sources of data are also discussed in this section.3.2 Theoretical FrameworkThe reviews of literature in chapter two reveal that there are basically four major hypotheses discussing the relationship between inflation and stock returns. These theories are1. Fisherian hypothesis2. procurator hypothesis,3. Tax-effect hypothesis and4. Inflation illusion hypothesis.The Fisherian hypothesis is thus specifiedWhere is the real returns, is the actual inflation which is the combination of the unexpected and expected inflation. While is the computer actus reus term that is distributed randomly and normally with zero mean and constant variance. This sign of determine if the specification is in line with the fisherian hypothesis. Thus a significant and positive sign suggest that stock hedges inflation while a negative sign suggest contrary.3.3 Model specificationBased on the outcome of our theoretical framework which attempts to explain the relationship between real stock returns and inflation, we specify the model for estimation. Stock return represented by all share indexes (ASI) is the dependent variable while the explanatory variables are, one-period lagged inflation represented by consumer indexes (CPI) and one-period lagged stock returns (ASI). This is based on the common smell that stock returns (ASI) takes some time to react to inflationary changes (CPI) and changes in all share indexes (ASI). In this study, it is assumed that stock returns depend on a set of variables denoted asTherefore, our empirical specification is stated as13.4 Methodology and Estimation ProceduresThis study makes use of Augmented Dickey laden (ADF) unit root test to check for the stationarity of the serial used in this study, Engle and Johansen cointegration tests is used to confirm if the serial have long run relationship while causal long run relationship is determine using an Error correction Model (ECM) which will reveal both the short run and long run relationship between inflation (LOGCPI) and stock returns (LOGASI).3.4.1 Unit Root TestAssume we have the following AR (1) process (1)and is a white noise e rror term. We can see the above expression by subtracting from both sidesThus (2)In practice, instead of estimating equation 1, we estimate equation 2 and test the hypothesis that =0. If =0 then that is we have unit root meaning the time serial publication is non-stationary ( for unit root is non-stationary). Thus we can take the first variety of and regress on to see if () is zero or not in order to confirm if the series are stationary or not. infra the null, the estimation for is not distributed T-student, so the Dickey Fuller test is required. We use the Augmented Dickey Fuller (ADF) table to correct for possibility of the error term () been auto correlated. The ADF test is specified in the equation beneath3Where is a white noise Error Term.3.4.2 Co integration TestsTrended data can be regarded as potentially a major problem for empirical econometrics. Trends may give rise to spurious regression and uninterpretable t- statistics. The stack reality is that in economics most time series are subject to some type of trend while differencing in series until it becomes stationary is one major solution. This has been shown that differencing can lead to handout of long run properties of a series. Based on this the combination of series that are difference once I(1) will give us a model that is stationary I(0). In achieving this aim this study consider two different co integration tests which are Engle and Granger co integration test and Johansen co integration test. According to Engle and Granger (1987), a time series and are give tongue to to be co integrated of order db where d b 0 written as CI (db) ifBoth series are integrated of order dThere exists a linear combination of these variables say which is integrated of order d-b. The vector and is called a co integrating vector.The Engle and Granger co integration test involve two steps the first step is conducting an OLS regression on the variables in the model specification. The second step is to condu ct an ADF test on the residual from the regression if the residual is stationary, then the series are said to be co integrated.The Johansen co integration test on the other hand involves the use of a VAR model and the different level best likelihood ratios are used to determine the co integrating vectors. These tests are trace test and maximum eigen value test. Different information criteria such as Akaike Information Criterion, Schwarz information criteria (SIC), Hannan-Quinn Information Criterion, Final Prediction Error and Sequential Modified test Statistic are used in determining the lag length.3.4.3 Error Correction ModelCo integration analysis provides a test for spurious correlation. Finding co integration between apparently correlated I(1) series validate the regression but failure to find co integration is an indication that spurious correlation mayhap present thus invalidating the inferences drawn from such correlation. Co integration analysis also helps in formulating th e process of dynamic adjustment. However time series data lose their long run properties when they are differenced allowing only for conclusions on the short run determinations. Therefore there is a need to construct a model that would combine both the short run and long run properties of the variables in the model. As suggested by Engle-Granger representation theorem that if two series are co integrated then they will be efficiently represented by an error correction mechanism. The Error Correction Model is used to capture both the short run and long run properties of the series. The method involves developing a model from it generalized form (over parameterized) to a specific form (parsimonious). In addition if the series are co integrated these dynamic specifications will encompass any other partial adjustment model. The error correction of the Auto regressive distributed lag (ADL) takes the form where the long run properties are derived from the proportionality between and. The above specification relates the short run change in the dependent variable to the short run change in the explanatory variable.this is called the impact effect () but ties the change to the long run impact through a feed-back mechanism.3.5 DataThe study will utilize monthly time series data from 19972006. Data for the variables will be sourced from Central Bank of Nigeria Statistical Bulletin (2006) and the Nigerian Stock Exchange Annual Reports (2006). The variables of interest in this study are all in logs. These variables are consumer price indexes (CPI) as inflation series and all share indexes (ASI) as stock returns.CHAPTER quartetteSUMMARY OF EMPIRICAL RESULTSThe summary of the statistics used in this empirical study is presented in the appendix. As can be observed from the Table, (see pagexx) the mean value of stock returns is 9.359606 while inflation is 8.442205. It is also observed that both LOGCPI and LOGASI are positively skewed. The kurtosis value is positively low and J arque-Bera (J-B) statistic test value is relatively high. These suggest that the two series are skewed to the right.Figure1below depicts the graphical metaphors of the data that were used in this empirical analysis. The figure reveals that stock return witnessed significant increase within the period of this study.Figure 1 Graphical illustration of statistics used in the analysisTable 1 Stationarity Test publicationVariablesLevelsFirst DifferencesADF 1ADF 2ADF 1ADF 2LOGASI0.712327