Icebergs and Government Productivity

Imagine a room full of economic forecasters discussing the state of the economy and its future performance. Armed with reams of data, all are well versed on key economic indicators, and each wants to look smarter than the next.

So, let's ask them what the growth rate of U.S. labor productivity was last year.

"Oh, OOOH!" grunts John E. Conomist, waving his hand furiously. "That's easy—4.3 percent."

Good. But that number is just for the private business sector. Who knows what productivity was for the nation's public sector?

Now imagine the vacuum silence of an iceberg floating past you in the ocean. That is the sound of a blank stare from our forecasters. But they are hardly alone, for no one on the planet could answer that question with more than a guess, and it's indicative of how little we know about roughly one-fifth of the U.S. economy.

To many, it's a case of so-whaddu-I-care. Labor productivity—an economist's term for work efficiency—is not exactly conversation fare at cocktail parties ("Say, I heard your son increased his trash removal productivity by almost 8 percent—bravo!"). Insert "government" into that discussion and it's iceberg-city.

Public sector productivity might lack sex appeal—it won't, for example, put money directly back in your pocket like a tax cut. But over the long haul, it can get your tax dollar to go further, and possibly demand fewer of them in the future. In some ways, productivity's "need to know" applies more to the public than to the private sector, particularly if one looks beyond the impact of private sector productivity on tax collection forecasts.

Private companies are driven by competition, and, regardless of the average productivity rate, high-achieving companies should out-compete lower-achieving ones. Such a performance lever is largely missing from government, which could have dramatic long-term consequences.

Richard Murray is chief economist with the Swedish government and one of the world's foremost experts on government productivity—one source called him a "bottomless well of experience and wisdom in this field." Murray said there will always be demands for greater efficiency in the public sector. But virtually all developed countries are staring down a resource dilemma, and government productivity will have a lot to say about how serious the matter becomes.

"It seems at this moment of time [productivity is] more crucial than ever since there will be high demands on service provisions—from students, elderly, retired, nonworking—while at the same time the working population diminishes," Murray said. "This will call for even higher tax rates unless we can increase efficiency."

Let's put the matter in more concrete terms. Annual spending by local, state and federal governments in the United States exceeds $2.5 trillion. Now, a goodly portion of government spending is transfers to or for individuals (think Social Security and Medicare). While some government workers are needed to process and manage such payment systems, even a rapid improvement in government productivity would not have much impact on total spending in such areas.

But so-called discretionary spending—appropriations for things like defense, transportation and national parks—is also on the rise. At the federal level, nondefense discretionary spending has risen by almost $110 billion, or 5 percent annually, since 1991, according to the Congressional Budget Office. By comparison, annual inflation averaged about 2.8 percent during this period. Further down the government ladder, state spending increased 30 percent from 1988 to 1997 after adjusting for inflation—nearly twice the growth rate of the economy and more than three times that of personal income, according to the Urban Institute.

Yet despite the increased spending across the board, we know virtually nothing about the public sector's current productivity, a formula that only fuels the fire of negative government stereotypes. "A lot of [politicians and citizens] are hung up on stereotypes. ... We see this very simple assumption often that government isn't productive," said Marc Holzer, chair and professor of Public Administration at Rutgers University and executive director of the National Center for Public Productivity.

In fact, research shows that government hasn't fared too poorly in (admittedly dated) productivity comparisons with the private sector. But no data have been collected since 1994—right about the time productivity rates took off in the private sector—and none is expected in the foreseeable future because there is virtually no pressure to do so, and because the task itself is fraught with deep, though not unmanageable, sinkholes.

What's the big deal?

Productivity has always mattered, but has been grabbing more headlines of late because it accelerated significantly in the private sector starting in 1995, roughly doubling to better than 2 percent. As such, private sector productivity has been an important factor in the nation's strong economic growth, and today is tracked closely by everyone from the Federal Reserve to stock market analysts to budget prognosticators.

In a February speech, Federal Reserve Governor Edward Gramlich said, "Of all the statistics that analysts pore over, productivity growth is surely the most important in the long run." Productivity determines the long-run path of living standards as well as projections for tax surpluses and entitlement trust funds, he said, adding, "[w]ere national leaders able to pick one economic statistic to be favorable, they would surely pick growth in productivity."

Despite this seeming importance, there is currently "nothing going on" in the United States as far as tracking aggregate productivity data for government, according to Donald Fisk, an economist and productivity expert with the Bureau of Labor Statistics (BLS). That wasn't always the case. In the early 1970s, there were a number of colliding anxieties over inflation, international competition with Europe and Japan, and rapid growth of government employment from Great Society programs of the 1960s.

"There was great concern at that point over government productivity and productivity in general," Fisk said. A snapshot report on the issue by the General Accounting Office (GAO) springboarded into an ongoing measurement program in 1973 whose "goal was to improve federal productivity," said Fisk, who was intimately involved with the effort.

At its height, the program collected productivity data on roughly two-thirds of all federal employees, and also tracked a sample of state and local government activities. At the federal level, data were just one leg of a larger, strategic program for improvement that included workshops, department reports and workplace consultants. Every agency also had its own productivity coordinator who was responsible for gathering and submitting output data for the centralized report, Fisk said.

Going postal: USPS delivering productivity Maybe letter carriers know someone is watching, and maybe that's why they coined the mantra of "neither snow nor rain ... " to describe mail delivery by the U.S. Postal Service. Although it's a favorite whipping horse for anti-government types, no government entity embraces the concept of productivity like USPS. A few other federal or quasi-federal agencies have it on their radar. The Federal Reserve System, for example, currently collects data on its payment services and is investigating a more formal, comprehensive program that would tabulate productivity for the payment services for each of the Fed's 12 district banks. But only USPS has a full-fledged tracking program on annual labor productivity, as well as total factor productivity. The reason is simple—the post office is required by law to be self-sufficient, according to Richard Strasser, USPS executive vice president and chief financial officer. "We have our own built-in motivation." From 1990 to 2000, labor productivity at USPS increased by 1.2 percent per year, on average, according to USPS figures. Three years-1990, 1993 and 2000—saw productivity grow between 2 percent and 4.4 percent, representing the lion's share of overall growth during this time. The productivity focus at USPS is on the organization, not individual workers. "There's not a lot [for individual workers] to do to work harder or faster," Strasser said. Rather, productivity growth typically come through systemwide improvements in processes and technology. But being on the plus side of productivity isn't always enough for success, either, even for a quasi-government monopoly. In some ways, USPS is an example of the futility of such measures for government, because there are extenuating circumstances not encountered in the private market. USPS, a $68 billion operation, expects to see a loss of at least $2 billion this fiscal year, Strasser said, and will seek a rate increase from the independent Postal Rate Commission, piggybacking one implemented just last year. It has also suggested cutting Saturday delivery. The problem, at least in part, is one of price inflexibility, Strasser said. "We don't have the ability to raise rates" to cover delivery costs. Under universal service policy, USPS is required to deliver to all addresses at uniform and affordable costs. But every year, close to 2 million new delivery destinations are added, which requires about 4,800 new letter carriers. E-mail and other competitors have also cut into the volume of letters and packages, making individual deliveries less cost effective. The rate increase also does not reflect service quality improvements at USPS. For example, on-time delivery for overnight and two-day items have both increased by better than 10 percent since 1995. "Just adding delivery points year in and year out and managing existing efficiency is a challenge," Strasser said.

But the program eroded over time. It got a big push out the door in the early 1980s when the Office of Management and Budget—the project's lead agency—decided to get out of the game, or rather, had it decided for them by newly elected President Ronald Reagan. "[President] Reagan came in and everything was dismantled," said Holzer of Rutgers.

Other agencies followed the OMB out the door, and in 1983 BLS was asked to carry the torch for the program, but efforts consisted of "just data" for the next decade, Fisk said. "There was no formal program to improve productivity and agencies lost interest." In 1994, spurred by budget cutbacks and lack of agency interest in productivity results, the BLS decided to shutter the effort altogether and "work on other things," Fisk said.

Some might be surprised to find that government productivity, though the data are old and rife with caveats, compared fairly well with those of the private sector. In its last official report, BLS found that annual productivity in the federal government grew by 1.1 percent from 1967 to 1994—not too far from the 1.4 percent annual growth of the nonfarm business sector during the same period. In fact, the two rates were identical from 1967 to 1982. But the rates diverged at this point, and from 1982 to 1994, the federal government's annual rate of productivity growth was less than half that of nonfarm businesses (0.6 percent to 1.3 percent, respectively).

But a better apple-to-apple comparison rightly involves the service sector of the private market, because most of what government provides are services. Here, government productivity looks pretty good because productivity for the private service sector actually declined by an average of 1 percent annually from 1989 to 1995, according the Council of Economic Advisers (though many believe measurement difficulties account for this anomaly, as discussed below).

A deadbeat, but in good company

Small consolation in this matter might come from the fact that the United States is hardly alone. Fisk said he knew of "very little active research" going on in any country. In fact, many countries don't even collect productivity data on the private sector, to say nothing of their government complement.

Sweden and Australia appear to be the only countries dedicating significant attention to the matter. Sweden has been measuring public sector productivity since at least 1960, covering between 60 percent and 70 percent of both federal and local government. Australia, on the other hand, is in the beginning stages of establishing a governmentwide measurement program, according to several government sources there.

But make no mistake, tracking government productivity is easier said than done. Before you can measure productivity, you have to first identify the outputs. Where there are physical goods or outputs, productivity measurement is straightforward. Think widget manufacturing—count the annual number of widgets produced, the hours it took to produce them, and then compare to the previous year, and, open sesame, you've unlocked the door to productivity measurement.

But defining government output—what we "get" for the tax dollars spent—is a morass, a Rubik's cube with a vague color scheme. "Measuring the output of the government sector is not so straightforward, as most of the things that governments provide are services," said George Verikios, senior research economist with the Australian Productivity Commission. "Services, by their very nature, can be difficult to measure. For example, how does one measure the output of the defense sector, the judicial system, welfare programs [or] education?"

There have been past efforts, Verikios acknowledged, but national statistical agencies in most countries "have done a poor job" of measuring public output. For example, the Australian Bureau of Statistics often used inputs like total spending to measure changes in output. "This assumes no change in productivity over time. Hence, only poor information on the productivity of services sectors, particularly government services, has been available," Verikios said. One of the main reasons BLS got out of the government measurement program was because it felt that even its private sector statistics were "sketchy, at best," according to Fisk. Rather than devote time and resources to reports no one was paying attention to, "it was better to get a handle on private sector productivity."

Some problems are unique to government services. Renowned Yale economist William Nordhaus said measuring productivity typically occurs "at the molecular level with some existing dollar or financial measure of an entity's output." Among other problems, he said, "most government goods and services are not priced. By this measure, then, conventionally measured output is zero."

But much larger problems plague the measurement of services in general. "We have a tough time measuring service output whether it's in the private or public sector," Fisk said. A recent article by Edward Dean of the BLS noted that "for a surprisingly large number of service-producing industries, there is a lack of agreement among economists on the best definition of output." In fact, he wrote, it was easier "to find fault with the output definitions currently used than to specify [correct] definitions."

Still other measurement difficulties exist, like the fact that current models do a poor job of capturing service improvements. For example, would you rather wait in line to renew your driver's license, or do so online? Simply counting the number of renewed licenses as annual output would ignore the convenience factor and resulting customer satisfaction realized by cyber renewal. External factors can also heavily influence activity-based output measures—like the effect of falling crime rates on police bookings and court cases.

These factors make it difficult to gather much enthusiasm for measuring government productivity. Because early measurement efforts would likely produce low-confidence data, it's simply easier to not start than to grope and plod a path to valuable and accurate data.

Hard, but not impossible

Even the Swedish model—the most rigorous effort in place today—is not without its uncertainties. An exhaustive 1996 report acknowledged that "previous measures of productivity aroused considerable interest and met with extensive criticism." Work has been done to incorporate quality improvements and other factors, but the report concludes that deficiencies were still present and results required further interpretation. "Shortcomings in adjustments for quality," the report stated, "are not unique to these [government-based] studies. Estimates of productivity within the private sector ... suffer from the same difficulties."

Most sources believe the methodological obstacles can be overcome, but only through sufficient effort and political backing. While the basic research issues are complex, "governments, by way of their national statistical agencies, have not devoted adequate resources to this problem," Verikios said. "From a political economy perspective, it is not hard to see why, as information on productivity can be used to pressure governments to improve productivity or reverse declining productivity in politically sensitive areas. This does not mean that the issue of measuring output for difficult-to-measure services is intractable. It is not. It's just that it has not been given adequate attention."

For its part, the BLS has been working with employers with "touchy-feely" jobs like architects, lawyers and others in the service industry to better define outputs, Fisk said, calling it "the first step of this very long road."

In fact, the private market now places more emphasis on so-called "total factor productivity"—a more complicated measure that considers innovation, capital and other factors to assess organizational productivity and success more comprehensively. But if tracking public sector labor productivity is government walking, total factor productivity is government running—to Mars and back.

For one, calculating total factor productivity requires good information on costs. The GAO found that nearly 40 percent of federal agencies did not receive unqualified or "clean" opinions on their 1999 financial statements, including the Department of Defense, which receives the lion's share of discretionary spending.

Reform flavor of the day?

Government watchers know that productivity is but another spoke in the wheel of government reform. It is, in some ways, subsumed by a broader current effort called performance measurement, which is intended to gauge social and other benefits brought about by government programs and spending. The movement got its federal legislative roots in 1993, when Congress passed the Government Performance and Results Act. GPRA was designed to shift government's focus away from activities, processes and inputs—like money spent, or staffing levels—and toward results.

One way to think about the two efforts is that productivity focuses on supply factors—the relationship between inputs and outputs for a service or product. Performance measurement, on the other hand, is preoccupied first with defining demand—who's receiving what benefit from a government program—and then measuring how well an agency performs in meeting that demand.

Put another way, performance measurement looks more toward outcomes of government spending—whether and to what degree intended results were achieved. Productivity looks instead to workplace efficiency and innovation, and to a large degree assumes that the work objectives, activities and results are properly aligned.

In some ways, it's a chicken-or-egg scenario. Productivity does indeed matter little if the resulting outcome has no impact or benefit, with the possible exception of saving marginal sums from the pyres of government spending. "I don't care how fast a government worker goes through a pile of paper until I know whether the pile of paper needs going through" in the first place, according to Jerry Ellig, senior research fellow at Mercatus Center at George Mason University. Productivity numbers "don't tell me a thing until I have a measure of the benefit" of government activity and spending.

GPRA is intended to fill that gap, and after six years of drafting missions and goals, two dozen major agencies began reporting progress in 1999. In its review of those annual reports, the GAO pointed out that agencies had limited abilities "to produce credible program performance and cost data" that would help identify improvement opportunities. Such limitations "were long-standing, and they will not be quickly or easily resolved."

A review by Mercatus Center found similar weaknesses regarding the availability and reliability of cost and other data. Agencies averaged a score of 31 out of 60 possible points for their GPRA reporting. Nonetheless, while GPRA has its critics, "it's the best lever we have [for government improvement]. It's what's there," Ellig said.

Holzer of Rutgers said that while measuring productivity was important, "you don't want to be hung up on the industrial definition of efficiency." Other measures—like service quality and knowing whether the services have their intended impact or benefit—were equally important in grading government's use of tax dollars. "We want to allow for a broader definition of what government's achieving."

But the very possibility (even likelihood) that government can spend trillions of dollars and still be unsure whether it is achieving intended—and also useful—outcomes might lie at the heart of the public's discontent with government. Murray, of Sweden, said, "Why is [productivity] not considered crucial in the public sector? Because of limited accountability. There will be no losses, nor any bankruptcy, just increased taxes."

Simply knowing annual productivity won't necessarily get you very far, as the previous federal attempt to measure productivity demonstrated. Nordhaus, the Yale economist, said that the key is how the data are used. "If you're not taking output numbers seriously, then it's not going to matter" what productivity levels are.

Holzer believed that tracking public sector productivity could put government on the path to greater public confidence, and act as a reminder "that the public sector is important and is adopting various innovations." But for any real changes to take root, "it has to come from the public and from politicians," Holzer said. If pressure to track government performance and pursue change "doesn't funnel down to politicians [from citizens] it's not going to happen."

Top