94

THE  PROBLEM  OF  THE  UNEMPLOYED

balance  periods  of  under-employment  and  over-employment
against  one  another.
Now  I  think  this  conclusion  is  not  well  founded.  The
following  line  of  illustration  will,  I  think,  make  it  apparent
that  a  net  loss  of  employment  over  a  long  period  is
caused  by  under  consumption  and  not  merely  a  maldistribution ­
  of  employment.
In  order  to  test  the  case,  take  a  community  with  stable
population  where  there  has  existed  a  right  economic  relation ­
  between  forms  of  capital  and  rate  of  consumption.
Suppose  an  attempt  is  initiated  to  increase  saving  by
abstention  from  consumption  of  some  class  of  goods,  say,
Cotton.  Increased  saving  has  no  legitimate  economic
channel,  if  we  suppose  this  parsimonious  policy  is  to
continue  and  not  to  be  balanced  by  an  early  period  of
abnormal  spending.  Since  no  trade  requires  increase  of
capital,  the  new  savings  may  as  well  be  invested  in  the
form  of  new  cotton  mills  as  in  any  other  way.  Let  us
suppose  that  the  over-saving  of  the  first  year  is  capitalised
in  this  form.  What  has  occurred  during  this  first  year
is  that  an  increased  employment  of  capital  and  labour
in  making  cotton  mills  has  balanced  a  diminished  employment ­
  in  making  cotton  goods.  Assuming  an  absolute
fluidity  of  capital  and  labour,  the  net  employment  for  the
community  is  not  affected  by  the  change.  People  have
simply  been  paid  to  make  cotton  mills  instead  of  to  make
cotton  goods.  At  the  end  of  the  year  there  exists  an
excess  of  cotton  mills  over  what  would  have  been  required
if  consumption  of  cotton  goods  had  stood  firm,  a  double
excess  over  what  is  needed  to  supply  the  now  reduced
demand  for  cotton  goods.  If  it  seems  unfair  to  anyone
that  I  should  apply  the  over-saving  to  the  only  trade