“We have 2 classes of forecasters: Those who don’t know… and those who don’t know they don’t know.” — John Kenneth Galbraith
I’ve been making a fortune lately. (No, I don’t own any Google IPO shares). Each month, I’ve been betting on the outcome of the Non-Farm Payroll report against my economist colleagues. I’ve been taking “the under,” and, over the past year, it’s been money 87% of the time. I expect this wager on a monthly jobs shortfall to remain successful for the foreseeable future.
Less lucrative, but much more fascinating than my book-making activity is the perplexing question “Why?” Why have the dismal scientists been unable to accurately discern what the employment situation is? It has certainly been perilous predicting job growth this business cycle; aside from a tendency towards over-optimism, what explains the consistent forecasting errors? Job growth predictions have been wronger, longer, and by a greater amount, than at any other time in the modern era of
This is an intriguing “whodunit” to me.
Nonfarm Payrolls, Post Recession: 2001-05 vs. Average Recovery
As Yogi Berra so wisely observed, “It’s tough to make predictions, especially about the future.” Those of us who work in glass houses – strategists, economists and weatherman – ought to be careful about throwing stones. But my crowd (Market Strategists) are typically wrong about the future. This cycle, Economists have
been unusually bad at predicting what happened just last month. The monthly consensus on Non-Farm Payrolls plays out like an old joke: “There are 3 types
of economists: Those who can count, and those who can’t.”
Clearly, something is amiss.
But rather than merely poking fun, we should be asking ourselves why this recovery is generating such weak job creation and correspondingly bad forecasts. Has something changed structurally? Are some basic assumptions about the business cycle flawed?
Perhaps econometric models are missing or over-weighting a key factor. Indeed,
what is it that nearly the entire field of economics has been somehow getting
I’ve been pondering this question for some time now. I have considered – and disposed of – the myriad excuses proffered: The disproved claims of the BLS Payroll Survey undercounting jobs versus their Household Survey; the uncounted “self-employed,
work-at-home-independent contractor;” that the Bureau of Labor Statistics
data is somehow bad; the rationale that (somehow) eBay is the explanation for 7 million missing jobs.
As a person unburdened by a Classical Economics education – I’m not an economist, but I sometimes play one on TV – I am free to ask the questions most economists can’t. I have my
suspects in the mystery of the awful economist. These are the most likely factors contributing to forecasting errors:
1. Globalization & Outsourcing
2. Productivity Gains
3. Post-Bubble Excess Capacity
4. ADCS (ERP) (Accelerated Depreciation)
5. Dividend Tax Cuts
6. Political Bias
7. NILFs (Not-in-Labor-Force)
8. Permanent versus Temporary Layoffs
10. Shell Shocked Executives
The first two points – Outsourcing issues and Productivity improvements – have been pretty thoroughly reviewed by economists – so neither of those issues is likely the cause.
But that still leaves a long list of unconventional issues that may be at least partly responsible for anemic jobs numbers . . .
UPDATE: March 5, 2005 7:25am
You can download the full report here.