Fuzzy Math in D.C.
by Paul A. Strassmann
strassmann.com
Computerworld
January 8, 2001
I first learned of the federal government's "hedonic" evaluations of IT budgets during budget reviews in 1992, when a Pentagon financial examiner noted that the Defense Department's requests for 5% annual increases in IT spending would actually be 22.8% in the first year and much more in following years. Such large jumps, the analyst suggested, would be unacceptable.
I checked the math. It was numerically correct. The large gains were generated on the basis of tables from the U.S. Department of Commerce's Bureau of Economic Analysis, so that the 5% for 1993 became 22.8% after adding a 17.8% implied annual cost reduction. The White House evaluated IT spending using "inflation-adjusted" comparisons. So a proposal for 5% salary increases would be OK because it matched the government's official "inflationary index." But proposals for computers were different, using a "hedonic deflationary index." This would boost the worth of computer hardware above its cash costs using the numbers indicated on the chart below.
Hedonic, derived from the Greek language, means "of or pertaining to pleasure." My budget examiners told me that "hedonic" was an economist's way of saying that customers would be deriving more pleasure from equipment that's better and cheaper to purchase. Checking the supporting data, I found that government economists used a mix of declining wholesale prices for desktops and laptops, plus unit costs of disk memories and printer performance statistics, to come up with indicators that reflected the decreasing cost and increased performance of hardware. So a 700-MHz desktop costing $3,000 would actually be twice as good as a 350-MHz desktop purchased for the same amount.
To understand how a hedonic deflator works, consider its consequences. Take a budget request of $1 billion in 1997 and $1 billion in 1998. With the aid of hedonic deflators, Department of Commerce economists would say that the $1 billion is actually worth $1.41 billion ($1 billion divided by 0.71) in 1997 and $2.13 billion ($1 billion divided by 0.469) in 1998, compared with the $1 billion spent in 1996.
There's no way an IT executive can respond to such logic, because it's flawed. Theoretical calculations of improved computer performance don't translate into IT budgets. Moore's Law works in electronics, but not in budgeting. Rising IT spending can't be judged simply by tracking catalog prices and hardware performance claims.
IT spending assessments that rely on deflationary cost indexing are used chiefly in the public sector, but this simplistic idea will certainly appeal to someone with an MBA, especially if he's told to find a rationale for cutting IT expenses. As IT budgets exceed 20% of the salaries of information workers, and total IT support costs exceed the median of U.S. manufacturers' profits, there will be more temptation to use government-approved hedonic deflators as a whip on the IT organization when the next recession hits. IT leaders must understand the fallacy of such measures before dealing with inquiries on why well-advertised cost reductions don't show up as decreased costs or bigger benefits.
Deflationary indexing isn't only a means for analyzing corporate budgets; it's also a clever way to manipulate whatever candidates claim in an election year. Recent boasts by the Clinton administration that IT has contributed one-third of U.S. economic growth are misleading. The numbers in the widely quoted Department of Commerce publication Digital Economy 2000 are biased because the IT sector's output was enlarged by hedonic indexing while the contributions of long established industries were shrunk because they were adjusted downward.
IT leaders should be careful in citing enormous nationwide productivity gains in case someone applies similar misleading metrics to their own firms. |
