[PEN-L:165] Re: FW: Rental equivalence in the CPI

["Eric Nilsson" <enilsson@xxxxxxxxxxxxxxx>]

(FYI, the below was also sent directly to Walter Lane)
-----------------------------------------------------
Re the message on rental equivalence approach from
Walter Lane of the BLS housing folks (who kindly
provided a quick response):

  > A lengthy article, which was published in the January 1983 CPI
  > Detailed Report explained the change in detail.

The explanation is clear.  But implication of the argument is that
the CPI is NOT intended to be a cost of living measure.

In short, the mandate to follow
    > Economic theory [which] suggested that the way to untangle the
    >consumption aspects from the investment aspects of housing
might be in conflict with the mandate to make the CPI a cost of
living measure.

Alternatively, while "economic theory" might suggest a way to
split the shelter and investment aspect of the costs of
homeownership, economic theory does not mandate that
the CPI should exclude the investment component of
homeownership. Rather, the exclusion by the CPI of
"investment components" of consumer expenditure is
accomplished _by definition_. That seems to be the
essence of the work done by Gillingham in reforming
the estimation of the costs of homeownership.

The switch to the rental equivalence approach was a step
away from the CPI being a cost of living measure.

That argument made by Boskin and others is that
   1) the CPI _should be_ a true cost of living measure
   2) therefore, we should eliminate all biases (defined
      narrowly as _upward_ biases)

But if #1 is accepted (Boskin et al accept it and so too does
the BLS) then what also follows is:
    3) the CPI should be supplemented by consideration of
       currently excluded components of consumer expenditure
       that are also relevant to the cost of living, such as the
       investment aspect of homeownership (including house
       insurance), interest expendences in general, along with
       income taxes.

Differentiating between the CPI and cost of living measure
is standard in the price index literature. That was really
my main point. But, economists who are not specialists in price
indices are not aware that an "unbiased" CPI would not be
a cost of living measure.

(Some say, of course, that a true cost of living measure should
include things like changes in the quality of life--increased
crime, etc--but these folks seem to accept that these non-material
things are the only difference between n unbiased CPI and a "true"
true cost of living measure. In fact, the CPI fails to be a cost of
living measure relevant to the standard of living, narrowly defined
as consumption of goods and services).

The cost of living is affected by both the prices of goods/services
related to current consumption and by expenditures for
items such as taxes, items such as interest costs on
credit cards, and items related to "investment" such as mortgage
rates, house insurance costs, and the like.

Such a differentiation between the CPI and a COL measure
is made by the exact same person who fronted the switch to the rental
equivalence approach, Robert Gillingham. (See Gillingham and
Greenless, "The Impact of Direct Taxes on the Cost of Living," JPE,
1987. See also Baye and Black, "The Microeconomic Foundations of
Measuring Bracket Creep and Other Tax Changes, Economic Inquiry,
1988).

Simple example of the above (ICH = investment cost of housing;
IAICH = income after ICH):

Year          1                           2
income = 10,000              11,000
CPI       =  100                      100
ICH      = 2,000                  4,000
IAICH   =8,000                   7,000

Here, income grew between year 1 and year 2 and the CPI remained
constant. By standard approaches we would claim that "real
income" grew.  BUT, the increased cost of the investment cost of
housing between the two years reduced money available to buy current
goods and services. Therefore, while "real" income apparently
grew over the two years, expenditures on current goods and
services actually fell. This is way items excluded from the
CPI must be considered when we try to determine changes
in the cost of living/standard of living narrowly defined
as current consumption.

More narrowly on the point of the specialness of the rise
in the cost of the investment aspect of homeownership:
   > The old method treated the purchase of a house as the purchase
   > of a consumer good.  This was at odds with the fact that a
   > house is actually an investment.  The index wound up treating
   > capital gains homeowner enjoyed as price increases that
   > consumers were suffering from.


The implication of the above is that an increased cost of the
investment aspect to one consumer (the buyer) is equal to the
capital gain to a second consumer (the seller). But,
many homes are new and the seller is a private firm.
And, the increased cost to the buyer is on average
less than the captial gain to the seller due to
costs such as monopoly rents to real estate agents, etc.
Further, if increased investment costs to the buyer is
cancelled out by increased income to the seller, then
why not take account of the fact that increased
prices of toasters might support increased income
for toaster workers? Finally, if the buyer and sellers
pay different interest rates on their mortgage, then
the costs to the buyer might be different from
the gain to the seller. (I'm sure with some assumption
of "perfect capital markets" some of  these problems will
disappear :>)   ).


According to a 3/3/97 Wall Street Journal article (page c15 of
the eastern edition):
   "[Stephen] Roach is especially bothered by
   the fact that the government relies, even in part,
   on a homeowner's guess of how much rent he (sic)
   could get, information he believes is 'alien to most
   families.'"

>From the same article:
  "Mr. Boskin recently acknowledged that the housing
   component is 'a problem right how.' On Friday, he
   told a bond market industry conference in Phoenix
   that the component was understated, although he declined
   to say how much."

Eric Nilsson
..
Eric Nilsson
Economics Department
CSUSB
San Bernardino, CA 92407
enilsson@xxxxxxxxxxxxxxx
909-880-5564