Those of us who live in the world of economic data swing between moods of elation and despair. The joy comes from the huge amount of high quality, timely, and consistent information that government statistical agencies turn out every week, creating puzzles and paradoxes for analysts to interpret and solve. The economic reports produced by the Federal statistical agencies and their partners in every state are the envy of the industrialized world.

But the despair comes from the wish that we had more. The need is particularly acute now, as the economy is changing course, creating new winners and losers among the nation's industries and among its regions.

The irony of economic statistics is that the quality of the data goes down whenever the geographic focus goes up. The Federal government, which can and often does run huge deficits, enjoys the most detailed, accurate data on the economy, while state governments like Indiana, who are prohibited from running deficits, are in the dark about crucial aspects of their economic performance.

The gap in quality is apparent in recent months from the employment data collected at the state, and the national level, as was recently reported by the consulting firm Economy.com. When you add up all of the employment estimates for the separate states and compare the total to the published statistics for the U.S. economy, you get a job total that is about 1 million on the high side. Although this discrepancy can be resolved in a number of ways, the most likely scenario is one where state employment totals are revised downward at some point in the coming months.

That might make Indiana's official job total, already down 35,000 jobs in July compared to a year ago in the preliminary counts, show an even bigger decline when better data arrive. Or, it may not. The point is that the state employment totals are subject to more uncertainty than we would like, given the fragile state of the economy and the state treasury.

What indicators are available for Indiana that are more reliable? That's a tough question, because all have their shortcomings. But here's a few that are always close to my computer keyboard.

One is average weekly hours in manufacturing facilities. These have been trending downward for more than two years, from a high point of almost 45 hours at the close of 1998, down to just over 40 hours in July. The speed of the declines, however, increased markedly at the close of last year, coincident with the national manufacturing slump. Average hours as a measure of manufacturing vitality is more robust than payroll counts, which are more sensitive to non-reporting.

Things get harder, however, when we try to assess the non-manufacturing side of the economy. One timely and interesting indicator is unemployment claims, published weekly by the Department of Workforce Development. They've been running consistently higher for the past eight months than their year-ago levels.

What's on my wish list? Tax receipts would be nice, especially for state sales taxes. Electricity and gas sales data have been interrupted by utility sales and mergers, but they could and should provide timely data on industrial activity. Until that day comes, we'll be tackling more mysteries than we should have to.