Page 141 - Contributed Paper Session (CPS) - Volume 2
P. 141

CPS1474 Jing R. et al.
                                                 |ℎ   ()| 2
                                      2
                                     , () =  ,
                                              ℎ , () × ℎ , ()
                The cross-spectrum technique is more intuitive and realistic in describing
             the correlation of CCCI and macro-economy indicators, and does not lose
             historical data information.

            3.  Result
                In our study, the data set includes the CCCI and its sub-indices, which
            span the period between 2009Q1 and 2018Q3. We use Real Estate Climate
            Index  (REC)  and  Shang  Hai  Interbank  Offered  Rate  (SHIOR)  to  represent
            investment status. And Total Retail Sales of Consumer Goods (RSCG) and CCI
            for domestic demand. Net Exports (NE) are utilized to on behalf of export
            condition.  We  also  use  the  GDP  as  indicators  of  general  economic
            development.  Except  CCCI,  other  variables  used  their  year-on-year
            fluctuation  ratio.  All  these  data  are  adjusted  to  quarterly  coinciding  with
            quarterly announcements CCCI by geometric mean. CCCI and its sub-indices
            are from the survey of CCCI Research Institute (CUEB and CUFE), while all the
            remaining variables from China National Bureau of Statistics. Since we use
            the year-on-year data, seasonally adjusted are not made in this paper.

            Consumer Confidence in Chinese Mainland















                          Fig1. Sub-indices time-varying parameter for CCCI

                 The first step is to determine the time-varying relationships between
             each sub-index and CCCI, however, before the analysis they are necessary
             to stabilize. Unit roots were tested via the augmented Dickey-Fuller (ADF)
             test in this paper for each time series. At the same time, we also give the
             normality and autocorrelation tests for estimating the results of the state
             space model, which suggests that the estimated models pass the diagnostic
             tests at a 1% level of significance. It is evident from Fig 1 that, during most
             observation periods, consumer psychology on material living standards has
             the largest positive impact on CCCI, obviously, they are more concerned
             about living conditions than other aspects. The attitude of the employment
             situation has increased its impact on their confidence. Their perception on

                                                               130 | I S I   W S C   2 0 1 9
   136   137   138   139   140   141   142   143   144   145   146