The Role of Demographic and Psychological Differences in Future Financial and Economic Expectations

Transcription

1 Internatonal Journal of Economcs and Fnance; Vol. 6, No. 12; 2014 ISSN X E-ISSN Publshed by Canadan Center of Scence and Educaton The Role of Demographc and Psychologcal Dfferences n Future Fnancal and Economc Expectatons Mahmoud Qadan 1 & Ur Ben-Zon 1 1 School of Management, Western Gallee College, Akko, Israel Correspondence: Mahmoud Qadan, School of Management, Western Gallee College, Akko 24121, Israel. Tel: E-mal: mahmodq@wgall.ac.l Receved: September 10, 2014 Accepted: October 23, 2014 Onlne Publshed: November 25, 2014 do: /jef.v6n12p37 URL: Abstract In ths study, agents are asked to assess the movements of short and long-term economc and fnancal varables. These varables nclude the prce level of goods and servces, real economc and fnancal varables. Dfferently from pror research, ths paper nvestgates short and long run expectatons, and tests the nfluence of racal composton of populaton on future prce expectatons. We analyze the agent expectatons from further aspects such as age, martal status, workng, psychologcal mood, ncome and wealth. The fndngs ndcate that expectatons are dfferent between demographc groups. Partcularly, females tend to expect hgher changes almost n the entre tested varables; ethnc mnortes systematcally report hgher expectatons; martal status has no mpact on future expectatons; and workng agents tend to be more conservatve regardng future changes. In addton, pessmstc agents generally tend to report hgher future prce movements manly for gold, nterest rates, unemployment and nflaton rates. Keywords: forecastng, nflaton expectatons, perceptons of specfc prces, survey data 1. Introducton Prce expectaton s consdered one of the crtcal varables n an economy. Snce the md-1990s, many central banks have targeted prce stablty as the prmary goal of ther monetary polcy, as a matter of fact that prce and prce expectatons have mportant and drect mpact on product and fnancal markets. Many economc authortes use to collect data from busness and consumer surveys that address questons on the movement of short-term economc varables such as output, nflaton and employment. The fnancal economc lterature emphaszes that forecastng nflaton expectatons based on consumer surveys have some superorty upon professonal and model-based forecasts (see e.g., Ang, Bekaert, & We, 2007; Hafer & Hen, 1985; Thomas, 1999). Therefore, Practtoners hghly apprecate ths knd of data, and even treat surveys results as a leadng ndcator of the economc actvty. Pror works have ntensvely studed the ssue of nflaton expectatons va studyng the ndexed-to-nflaton and non-ndexed bonds. Other works tested the accuracy and the ratonalty of varous survey measures of nflaton expectatons. These measures, generally, have a predctve power and ndcate a good ft wth observable nflaton. Snce agents n economy have to make a varety of economc actons such as consumpton, savng, nvestng and borrowng, the ssue of expectng economc varables - such as nflaton, nterest rates on deposts or loans, stock market return and economc growth- becomes crucal. Whle pror works manly concentrate on the ssue of nflaton expectatons, ths study offers a snapshot on ndvduals short and long-run expectatons regardng three man varables: goods and servces (ol, gold, electrcty); real-economc varables: GDP, nflaton and unemployment); fnancal varables (local stock market ndex; foregn market ndex; nterest rates on: deposts, mortgages, nter-bank nterest; exchange rates aganst the dollar and the euro; housng prces and other varables). Ths research dffers from pror research n several aspects: whle pror works focus on studyng nflaton only, ths artcle expands the varables of nterest to three categores: goods and servces, real-economc varables and fnancal varable. Second, the sample conssts of mult-cultural respondents. Thrd, respondents are all college students that allow controllng the effect of fnancal lteracy. Forth, respondents psychologcal mood s also 37

2 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 consdered n order to grasp ts mpact on ndvduals future expectatons. Ths type of data allows us to examne whether varous demographc dfferences have an mpact on the perceptons of the expected changes n economy varables. For ths purpose, we collected (meanwhle) data from 261 questonnares. In each questonnare, ndvduals are asked to provde ther one-year and fve-year expectaton wth respect to the movements of the three types of varables lsted above. The results, by and large, ndcate that future expectatons are dfferent between agents, and that these dfferences can stem from factors such as mood, beng part of mnorty, gender, and beng employed. More specfcally, males tend to be more conservatve regardng future prce movements, workng ndvduals have lower expectatons about future movements of real-economc and fnancal varable rather than non-employed, psychologcal mood has also an mportant role n formattng expectatons,.e., pessmsm have a postve mpact on the assessment of future prces. Pessmstc respondents tend to expect hgher prces for gold and ol compared to optmstc respondents. Fnally, belongng to mnorty feature pushes agents to provde very volatle and hgher future expectaton. Other factors such as wealth, age, degree of relgosty, martal status, ncome, and expected ncome have no clear mpact on the ndvdual future prce expectatons. The remander of ths study s organzed as follows. Secton 2 revews the relevant lterature. Secton 3 descrbes the data and develops the method and the econometrc model. Secton 4 dscusses the fndngs and the emprcal mplcatons, and the last secton provdes a summary and concludng remarks. 2. Lterature Revew There s an enhanced prolferaton n research that deal wth the queston whether agents n economy are able to detect early sgnals about future rates of economc growth. Future economc expectatons are manly studed on the base of two research streams. The frst s bult upon consumer confdence (CC) measures whle the second s bult upon model-based forecast of standard macro economc varables. Recent academc lterature has seen a rse n studes nvestgatng the effect of CC on stock returns. Actually, measures of consumer confdence have got ncreasng emphass as a measure of montorng consumpton trends. Many works have addressed emprcal questons on whether CC s nformatve of current or future consumpton, and whether CC has any predctve power over and above standard macroeconomc varables. The research shows that CC s a concdent ndcator of consumpton growth and ncome growth as well as changes n housng wealth, nflaton, unemployment and real nterest rates (Acemoglu & Scot, 1994). In addton, the hypothess that CC does not cause GNP (n the Granger sense) s rejected (Matsusaka & Sbordone, 1995). In addton, CC s Granger-Caused wth changes n the captal market (Jansen & Nahus, 2003). Measures of CC do forecast future changes n labor earnngs and non-stock market wealth, but measures of consumer atttudes appear to be drectly related to future consumpton growth, not just ndrectly through ther predctve power for household ncome or wealth (Ludvgson, 2004). When CC s hgh, future stock returns tend to be lower and vce versa, and the mpact of CC on returns s stronger for countres that have less well developed market nsttutons and for countres that are culturally more prone to nvestor overreacton (Schmelng, 2009). Pror research fnd also that CC measure have a predctve ablty wth respect to stock market returns (e.g., (Fsher & Statman, 2003; Brown & Clff, 2005) and even more pronouncedly for frms that are hard to prce and thus dffcult to arbtrage (e.g. Baker & Wurgler, 2006; Lemmon & Portnaguna, 2006). There are a negatve relatonshp between the CC and future stock return. In addton, there s appostve and statstcally sgnfcant correlaton between changes n CC and the contemporaneous stock returns, hgh stock return boost the CC (Fsher & Statman, 2003). De Brun et al. (2010) replcate pror results n the lterature (see e.g., Blanchflower & Colle, 2009; Lombardell & Saleheen, 2003) and fnd that hgher nflaton expectatons are reported by female, poorer, sngle and less educated. These observed demographc dfferences n nflaton expectatons were mostly explaned by varatons n age and educaton. They also fnd relatvely hgher nflaton expectatons among respondents who are nonwhte vs. whte, sngle vs. marred or lvng wth a partner, low vs. hgh ncome, at most hgh-school educated vs. more educated, and older. De Brun et al. (2011) allow respondents to report ther pont forecasts as well as ther densty forecasts for prce and wage nflaton. They fnd that those who express hgher levels of expectatons also tend to express hgher uncertanty n ther subjectve forecasts, and that heterogenety n expressed uncertanty s assocated wth demographc characterstcs and fnancal lrteracy. Gven the mportance of prce expectatons, ths study wll attempt to extend the lterature by addressng more varables than done before n major pror works, and to assess the mpact of demographc and psychologcal dfferences effect on ndvdual expectatons. 38

3 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; Data and Method Our sample conssts of 261 respondents from the Western Gallee College and the Unversty of Hafa, n Israel. All partcpants are students toward the frst degree who study n the departments of economcs, educaton and mult-dscplnary studes. Ths data selecton controls the problem of over-estmate nflaton of populaton groups that are wthout academc educaton (Mcormack et al., 2009). The age of the partcpants range from 18 to 54 (Mean=25.3, SD=5.51). The populaton n Israel s constructed of two man natonaltes: Jews, as the majorty and Arabs, as the mnorty. Our method conssts of t-tests and econometrcal model. The t-tests, ntensvely utlzed n the emprcal fnancal lterature, are employed to measure the averages of the varables. The econometrc model s based on a lnear regresson as follows: Y N = + =1 α β x + ε, (1) where y and x are the dependent and the ndependent varable, respectvely. As mentoned n the precedng sectons, the dependent varable s the ndvdual s one-year and fve-year expectaton on the rate of change of the level of the followng 13 varables: housng, ol, gold, electrcty, GDP, unemployment rato, the local captal market ndex, U.S captal market ndex, the nterest on deposts, mortgages, nter-bank nterest (prme nterest), and the exchange rate of dollar and euro currences aganst the local currency. The ndependent varables nclude the followng dummy varables: gender, martal status, beng employed, havng apartment as a proxy for wealth, and belongng to mnorty. The ndependent varables nclude also the followng scale varables: psychologcal mood and relgosty. In addton, we have the followng level varables: age and current ncome n local currency. For each dependent varable tested, the partcpant was nformed about the level prce n the day he fulflled the questonnare. Each partcpant was asked to provde hs future expectaton regard one year and fve years ahead. We then calculate the rate of change for each of the Y varables. 4. Emprcal Fndngs Table 1 presents some descrptve statstcs of the dummy varables. Accordng to ths table, the number of partcpants s 261, among of them 177 are females. Also, 92 of the partcpants are marred, and 171 have a job. Only 42 partcpants own an apartment, and the majorty of the partcpants are Jews (136) whle the rest are Arab. Table 1. Basc statstcs on the partcpants of the sample Gender Martal Status Work Apartment Mnorty/Majorty 177 Females 166 Bachelor 90 Do not work 42 Own 136 Jews 84 Males 92 Marred 171 Have a work 219 Do not own 125 Arab 3 Dvorced Note. Table 1 presents basc statstcs on the partcpants of the survey. Table 2 presents the descrptve statstcs of the one-year-ahead expected dependent varables, whle table 3 presents the fve-year-ahead expected dependent varables. In each table, the varables are categorzed nto three groups: goods and servces, real-economc, and fnancal varable. Accordng to these tables, the long run expectatons are larger than the expectatons of one year ahead. 39

4 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 Table 2. Descrptve statstcs of the dependent varables: one-year expectatons Panel A. Goods and Servces Real Economc Varables Ol Barrel Gold Ounce House Electrcty CPI GDP Unemployment average 7.50% 2.84% 6.41% 9.89% 1.74% 0.86% 2.14% stdev 12.08% 9.81% 12.81% 10.25% 3.20% 4.03% 7.51% max 60.38% 60.17% 43.48% 78.57% 18.77% 23.89% 7.40% mn % % % -8.05% % -2.00% medan 3.77% 2.20% 8.69% 7.14% 0.57% 0.88% 1% Panel B. Fnancal Varables $1 1 TA-100 S&P500 Mortgage Depost Prme nterest average 1.61% 0.20% 3.70% 24.43% 0.33% 0.66% 0.54% stdev 9.67% 11.95% 18.00% 17.43% 0.82% 1.19% 0.90% max 32.98% 40.28% 59.33% 43.54% 4.50% 8.00% 6.00% mn % % % % -2% -1% -1.25% medan 1.06% 0.20% 0.28% 33.97% 0.2% 0.2% 0.3% Note. Table 2 presents the descrptve statstcs of dependent varables that are related to one-year ahead expectatons. CPI resembles the consumer prce ndex, GDP s the gross domestc product, 1$ s the exchange rate of change of on Amercan dollar aganst the local currency, 1 s the exchange rate of change of one euro aganst the local currency, TA-100 s the leadng stock ndex n Tel-Avv stock exchange. Table 3. Descrptve statstcs of the dependent varables: fve-year expectaton Panel A. Goods and Servces Real Economc Varables Ol Barrel Gold Ounce House Electrcty CPI GDP Unemployment avg 21.45% 15.15% 14.73% 20.13% 7.17% 5.36% 4.83% stdev 26.53% 30.89% 18.07% 21.89% 10.05% 9.92% 10.02% max 88.68% % 47.83% 69.64% 57.47% 32.74% 7.40% mn % % % % -5.17% % -6.00% medan 22.64% 8.30% 17.39% % 4.42% 2% Panel B. Fnancal Varables $1 1 TA-100 S&P500 Mortgage Depost Prme nterest avg % 14.54% 29.99% 1.06% 1.71% 1.72% stdev 19.39% 19.04% 25.98% 21.61% 1.77% 2.23% 2.03% max 59.57% % % 62.68% 5.40% 14.00% 11.00% mn % % % % -4.50% -1.00% -3.50% medan 6.38% 6.21% 2.62% 35.64% 0.5% 1% 1% Note. Table 3 presents the descrptve statstcs of dependent varables that are related to one-year ahead expectatons. The notatons are as n Table 2. Table 4 presents the estmatons results of Eq. (1). Ths equaton has been estmated several tmes n order to explore specfcaton errors. That s, Eq. (1) ncluded n the frst estmaton all the canddate explanatory varables, whereas eventually the lst of good ft explanatory varables was reduced to only 4 varables: gender, belongng to mnorty, havng a work, and psychologcal mood level. That s, 40

5 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 Y = α + β Gender + β Mnorty + β Work + β Pessmsm + ε, (1) In each questonnare there are 9 questons that address the ssue of pessmsm. Those questons are formed as 1-to-5-scale questons n order to reflect the mood of ndvduals. For each questonnare we calculated the average of these nne questons, and the receved average value was employed as a proxy of ndvdual mood. Ths average value was utlzed as an explanatory varable. Table 4. Estmaton results of Eq. (1) Gender Mnorty Work Pessmsm R 2 F Goods and Servces Hosng: One-Year Expectaton *** *** *** Fve-Year Expectaton *** 0.146* *** *** Ol: One-Year Expectaton *** *** *** Fve-Year Expectaton *** *** *** Gold: One-Year Expectaton *** *** Fve-Year Expectaton *** 0.113*** *** Electrcty: One-Year Expectaton *** *** Fve-Year Expectaton * 0.119*** *** Real Economc Varables GDP: One-Year Expectaton Fve-Year Expectaton *** Unemployment: one year Expectaton *** Fve-Year Expectaton *** CPI: One-Year Expectaton *** *** *** Fve-Year Expectaton * *** *** Fnancal Varables TA100: One-Year Expectaton *** Fve-Year Expectaton *** SP500: One-Year Expectaton 0.098*** *** *** Fve-Year Expectaton *** 0.065*** *** Mortgage: One-Year Expectaton *** 0.283*** *** *** Fve-Year Expectaton *** 1.106*** *** *** Depost (nterest): One-Year Expectaton *** *** Fve-Year Expectaton *** *** *** Prme (nterest): One-Year Expectaton *** *** Fve-Year Expectaton *** 0.726*** * 0.688*** *** $ Exchange rate: One-Year Expectaton *** *** *** Fve-Year Expectaton *** *** Exchange rate: One-Year Expectaton *** *** *** Fve-Year Expectaton *** *** *** Note. Table 4 reports the estmaton results of Eq. (1). (***), (**) and (*) ndcate sgnfcance at the levels of 1% 5% and 10%, respectvely. Table 4 shows that β 1 coeffcent s generally negatve. β 1 coeffcent quantfes the mpact of gender on the expectatons of ndvdual. Of 26 values of the β 1 coeffcents, there are 21 negatve values. 12 of them are statstcally sgnfcant. Ths fndng mples that there s a sgnfcant dfference between the expectatons between males and females. More specfcally, males tend to lower ther expectaton regardng future movements n the varables tested (except for ther expectatons regard the S&P500). β 2 assesses the mpact of beng a mnorty member on the ndvdual s expectatons. Table 4 shows also that β 2 s manly postve. Of 26 values of the β 2 coeffcents, 23 are postvely sgned and only 4 are negatve. For 9 cases β 2 s statstcally sgnfcant. Ths fndng s n lne wth Bryan and Venkatu (2001) and Brun et al. (2010). They argue that racal ethnc mnortes (non-whte vs. whte partcpants) systematcally report hgher, and sometmes unrealstc, nflaton expectatons. Our fndngs show that mnorty s expectatons are hgher not only for nflaton expectatons, but also for other several economc and fnancal varables. We fnd also that ndvduals that have a work (job), and consequently have hgher ncome, tend to lower ther 41

6 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 future expectatons about movements n the economy varables, as appears n the negatve β 3 coeffcents as appears n Table 4. β 3 receves negatve values for 15 cases, whle for 11 cases the coeffcent s negatve. Ths fndng s n lne wth Zkmund et al. (1999), accordng to whch low-ncome populatons tend to be more myopc when makng fnancal decsons. Also, ndvduals who report hgher nflaton expectatons may have shorter fnancal plannng horzons. These populaton groups seems to be more senstve to transent prce shocks as well as less nformed about the longer-term prce trends captured n the nflaton rate (Brun, 2010). The proxy measure of ndvduals pessmsm tends to push the future expectatons up as reported by the β 4 coeffcents. The β 4 coeffcent estmates the mpact of pessmsm level on ndvdual s future expectatons. β 4 coeffcents are sgnfcantly postve n most cases. The nterpretaton of ths fndng can be related to other observed fact accordng to whch, pessmstc ndvduals provde hgher absolute changes. In addton, pessmsm s hghly lnked to the mnorty respondents. We compared between the psychologcal mood of majorty and mnorty respondents by the followng two equatons. Optmsm= θ 0 + θ 1 Mnorty, (2) (78.19) (-4.03) where optmsm s a scale varable calculated by takng the average of 7 questons that addressed optmsm, and mnorty s a dummy varable whch receves the value of 1 f the respondent belongs to mnorty n the populaton. The estmaton result shows that θ 1 s postve and statstcally sgnfcant, ndcatng that the optmsm level s lower for mnorty respondents. In addton, we run the followng equaton: Pessmsm =θ 0 + θ 1 Mnorty, (3) (39.83) (7.49) Where pessmsm s also a scale varable calculated by takng the average of 9 questons that address pessmsm. The results of Equatons 2 and 3 reveal an nterestng fndng. They ndcate that ndvduals belongng to mnorty are less optmstc, or statng dfferently, more pessmstc. The results n Table 4 are relatvely well explaned for case of the fnancal and for the goods varables. However, the real-economc varables are weakly explaned by the proposed explanatory varables. Another nsght on the dfferences between expectatons of majorty vs. mnorty s gven n table 5. The table provdes the t-test result on whether ndvduals that belong to mnorty have systematcally hgher values of future expectatons rather than the majorty. Table 6 reports the relatonshp between optmsm level of ndvduals and ther expectatons regardng movements n stock market ndex. The results are statstcally sgnfcant (for the S&P 500) and show that optmsm postvely affect the expectatons on movements n the captal market. Table 7 reports the correlaton between ndvduals pessmsm level and ther expectaton regardng movements n captal market ndex. We fnd that pessmsm has a negatve affect (as expected) on the future expected movements n the market ndex. Table 5. Dfferences n expectatons of majorty vs. mnorty ndvduals n (%) Majorty Mnortes T-test Value Unemployment one-year expectaton fve year-expectaton *** Prme Interest one-year expectaton *** fve year-expectaton *** Mortgage one-year expectaton *** fve year-expectaton *** One-Year Depost one-year expectaton *** fve year-expectaton ** Housng one-year expectaton *** fve year-expectaton *** Ol Barrel one-year expectaton ** fve year-expectaton *** Gold One Ounce one-year expectaton ** 42

7 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 fve year-expectaton ** Tel-Avv stock exchange ndex one-year expectaton fve year-expectaton S&P 500 ndex one-year expectaton fve year-expectaton CPI one-year expectaton fve year-expectaton Dollar Exchange rate one-year expectaton fve year-expectaton Euro Exchange Rate one-year expectaton fve year-expectaton *** GDP one-year expectaton fve year-expectaton *** Electrcty one KW/H one-year expectaton fve year-expectaton Note. Table 5 reports the dfferences of future expectaton of majorty versus mnorty ndvduals. (***), (**) and (*) ndcate sgnfcance at the levels of 1% 5% and 10%, respectvely. Table 6. Relatonshp between captal market ndex and ndvdual s optmsm Dependent Varable alpha beta R 2 F TA TA-100 (5-Year Forecast) S&P *** S&P 500 (5-Year Forecast) *** *** Note. Table 6 presents the estmaton results of the equaton Y = a + β Optmsm + ε. The beta coeffcent lnks the Optmsm measure to the future expectaton of ndvduals. TA-100 denotes the (domestc) Tel-Avv 100 stock ndex. (***), (**) and (*) ndcate sgnfcance at the levels of 1% 5% and 10%, respectvely. It s commonly argued n the lterature that durng a pessmstc perods (e.g., crses, wars, captal market crashes and hyper-nflaton) substantal speculatve captal flows are dverted to durable goods, among them, and maybe the most promnent of whch, s gold. Recent lterature relates part of the change n gold prce to psychologcal sentments. More specfcally, t s argued that gold prce s drven by the Volatlty Index (VIX) whch s wdely consdered as an ndcator of pessmsm sentment of the market partcpants (Qadan & Yagl, 2012). In addton, gold acts as a safe haven asset, offerng protecton to nvestors aganst losses n fnancal markets (Baur & McDermott, 2010). We test whether bad psychologcal mood of respondents forecast a postve change n gold prce. For ths sake, we run the followng regresson model: Gold = θ 0 + θ 1 Psmsm, (4) (-1.524) (3.319) where Gold s the long-run rate of change n the expected prce of gold provded by respondent, and Psmsm s the level of pessmsm of respondent. The estmaton result n Eq. (4) shows that θ 1 s postve and s statstcally sgnfcant, ndcatng that pessmsm level pushes gold prce upward. Ths fndng s n lne wth pror research accordng to whch market partcpants fear sentments about droppng prces n the captal market are assocated wth rsng n gold prce. Ths fndng may strengthen the clam that gold acts as a safe haven asset, offerng protecton to nvestors aganst losses n fnancal markets. 43

8 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 Table 7. Relatonshp between captal market ndex and ndvdual s pessmsm Dependent Varable alpha beta R 2 F TA TA-100 (5-Year Forecast) S&P *** *** *** S&P 500 (5-Year Forecast) 0.536*** * * Note. Table 7 presents the estmaton results of the followng equaton. The explanatory varable s the pessmsm measure. Hence, beta lnks the pessmsm sentments to the future expectatons. (***), (**) and (*) ndcate sgnfcance at the levels of 1% 5% and 10%, respectvely. Y = a + β Passmsm + ε. 5. Concludng Remarks It has been lengthly argued that expectatons about future nflaton are generally mportant for households decsons about spendng and savng, and for central bankers for calbratng monetary polcy. However, snce agents n economy make a varety of economc decsons about consumpton, savng, nvestng and borrowng, among other thngs, the ssue of ndvdual economc expectaton of other economc and fnancal varables becomes crucal. Ths paper offers an extenson to the ssue of ndvdual s prce expectatons. Whle pror works on economc expectatons focused manly on forecastng nflaton, lttle s known about forecastng movements of other mportant varables n the economy. Specfcally, nstead of focusng on one varable only (the nflaton), we extend the varables of nterest to three categores: goods and servces, real-economc varables and fnancal varables. These varables had been studed n the context of short and long-run. For ths purpose, we constructed questonnares that address the ssue ndvdual s expectatons wth respect to the varables lsted above. The fndngs ndcate that expectatons among ndvduals are dfferent, but can be descrbed as follows: males expectatons are more conservatve comparng wth females; workng respondents report lower prce expectaton relatve to non-workng respondents; Arab ndvduals (as the mnorty n the populaton) report systematcally hgher expectatons comparng wth the majorty of the populaton. The sgnfcant tendency of mnorty agents to hghly expect future changes can be related to other observed phenomenon accordng to whch mnorty respondents are more (less) pessmstc (optmstc) comparng to the majorty. In addton, pessmstc agents have a clear tendency to report hgher gold prce expectatons. Ths fndng s n lne wth the clam that gold can be substtute and safe haven nvestment vehcle n perods of nstablty. References Acemoglu, D., & Scott, A. (1994). Consumer confdence and ratonal expectatons: are agents' belefs consstent wth the theory? The Economc Journal, Ang, A., Bekaert, G., & We, M. (2007). Do macro varables, asset markets, or surveys forecast nflaton better? Journal of Monetary Economcs, 54(4), Baker, M., & Wurgler, J. (2006). Investor sentment and the cross secton of stock returns. The Journal of Fnance, 61(4), Baur, D. G., & McDermott, T. K. (2010). Is gold a safe haven? Internatonal evdence. Journal of Bankng & Fnance, 34(8), Blanchflower, D. G., & MacColle, C. (2009). The formaton of nflaton expectatons: an emprcal analyss for the UK (No. w15388). Natonal Bureau of Economc Research. Brown, G. W., & Clff, M. T. (2005). Investor Sentment and Asset Valuaton. The Journal of Busness, 78(2), Bryan, M. F., & Venkatu, G. (2001). The demographcs of nflaton opnon surveys. Federal Reserve Bank of Cleveland, Research Department. De Brun, W. B., Vanderklaauw, W., Downs, J. S., Fschhoff, B., Topa, G., & Armanter, O. (2010). Expectatons of nflaton: The role of demographc varables, expectaton formaton, and fnancal lteracy. Journal of Consumer Affars, 44(2), De Brun, W. B., Mansk, C. F., Topa, G., & Van der Klaauw, W. (2011). Measurng consumer uncertanty about future nflaton. Journal of Appled Econometrcs, 26(3),

9 Internatonal Journal of Economcs and Fnance Vol. 6, No. 12; 2014 Fsher, K. L., & Statman, M. (2003). Consumer confdence and stock returns. The Journal of Portfolo Management, 30(1), Hafer, R. W., & Hen, S. E. (1985). On the accuracy of tme-seres, nterest rate, and survey forecasts of nflaton. Journal of Busness, Jansen, W. J., & Nahus, N. J. (2003). The stock market and consumer confdence: European evdence. Economcs Letters, 79(1), Lemmon, M., & Portnaguna, E. (2006). Consumer confdence and asset prces: Some emprcal evdence. Revew of Fnancal Studes, 19(4), Lombardell, C., & Saleheen, J. (2003). Publc expectatons of UK nflaton. Bank of England Quarterly Bulletn, 43(3), Ludvgson, S. C. (2004). Consumer confdence and consumer spendng. Journal of Economc Perspectves, Matsusaka, J. G., & Sbordone, A. M. (1995). Consumer confdence and economc fluctuatons. Economc Inqury, 33(2), McCormack, L., Bann, C., Uhrg, J., Berkman, N., & Rudd, R. (2009). Health nsurance lteracy of older adults. Journal of Consumer Affars, 43(2), Qadan, M., & Yagl, J. (2012). Fear sentments and gold prce: testng causalty n-mean and n-varance. Appled Economc Letters, 19, Schmelng, M. (2009). Investor sentment and stock returns: Some nternatonal evdence. Journal of Emprcal Fnance, 16(3), Thomas, L. B. (1999). Survey measures of expected US nflaton. The Journal of Economc Perspectves, Zkmund-Fsher, B. J., & Parker, A. M. (1999). Demand for rent-to-own contracts: a behavoral economc explanaton. Journal of Economc Behavor & Organzaton, 38(2), Copyrghts Copyrght for ths artcle s retaned by the author(s), wth frst publcaton rghts granted to the journal. Ths s an open-access artcle dstrbuted under the terms and condtons of the Creatve Commons Attrbuton lcense ( 45