Abstract:

In  an  economy  that  applying  the  free  currency  rate  regime,  the  Exchange  rate  is  not  constant,   it   represents   the   fluctuations   with   time.   In   this   study,   the   analysis  completely   based   on   the   historical   Exchange   rate   data.   Therefore   Exchange   rate  data  can  be  considered  as  grey  value.  Traditional  grey  models  try  to  describe  the  evaluation   the   original   data.   But   those   models   are   not   a   suitable   tool   for   non-­‐stationary   random   data.   For   this   reason,   in   this   study,   we   use   the   grey   Markov  model.  A  Markov  chain  which  the  future  evaluation  only  depend  on  present  state  it’s  no  dependency  to  previous  states,  so  a  Markov  chain  can  be  used  to  model  the  unstable   systems.   The   future   values   of   system   can   be   predicted   using   the   grey  model.  In  the  grey  Markov  model  approximation,  firstly  Exchange  rate  estimated  by  GM(1,1)   model   and   the   error   is   obtained   as   difference   between   the   estimated  value  and  real  data.  This  new  series  is  divided  into  finite  state.  Then  observing  the  transients   between   the   states   Markov   probability   matrix   is   obtained.   The   future  values   time   series   are   predict   by   this   matrix.   Grey   Markov   model   which   is  constructed   by   GM   (1,1)   model   and   Markov   chain,   allow   us   to   make     the   beter  estimation  for  exchange  rate.  In  this  study,  daily  TL/USD  Exchange  rate  data  is  used.  The   data   is   collected   from   TCMB.   Finally   it   has   determined   that   the   grey   Markov  model  is  a  good  approximation  to  predict  the  currency  Exchange  rate.     GRİ  MARKOV  KESTİRİM  MODELİ  KULLANILARAK  DÖVİZ  KURU  TAHMİNİ

Keywords: