INSIGHTS: MAKE COST BENEFIT ANALYSIS MEANINGFUL
Guest Columnists: James Broughel, Senior Research Fellow, Mercatus Center at George Mason University Reprinted with permission from The Regulatory Review.
In 1947, economist Tjalling Koopmans wrote a review of the book Measuring Business Cycles,written by his fellow economists Arthur Burns and Wesley Mitchell. The title of Koopmans’s review, Measurement Without Theory, referenced the fact that Burns and Mitchell had collected and summarized impressive amounts of macroeconomic data and described in detail the business cycle fluctuations they observed, but they had not drawn sufficiently on economic theory to explain the observed regularities in their data. By omitting theory that might otherwise give their data context and meaning, Burns and Mitchell’s contribution was cast into doubt.
In more recent years, there is another area where theory has taken a backseat to empirical measurement: Namely, the theory behind benefit-cost analysis (BCA). BCA intends to predict how regulations and other public policies impact society for better or worse. Conspicuously missing from BCA, however, is the necessary theory connecting whatever BCA is measuring to the well-being of actual people.
Debates over the relative merits of theory versus empirical measurement are not new to academic inquiry. In the late 19th century for example, a debate took place in the German-speaking world over the importance of theory as it pertains to economics. On one side of the Methodenstreit, as the debate came to be known, was the so-called German historical school, led by Gustav Schmoller. Schmoller cautioned against fundamental laws in economics, arguing that historical context and culture are always changing and that outcomes will vary by time and by place. Theory can be developed, he accepted, but careful measurement comes first. Only after extensive data collection and analysis can hard conclusions be drawn, or so members of the historical school believed.
A group of economists in Austria, led by Carl Menger, viewed matters differently. They saw an important role for theorizing in the social sciences, believing that a scholar can construct an idealized model economy and, through careful reasoning, arrive at certain fundamental truths about its workings. Insights can then be extended to the messy and complicated world around us, with its nearly unlimited and imperfect data that can often be contradictory and misleading. Furthermore, any analysis of data involves theory—theory is impossible to escape—because understanding what data represent is critical to their interpretation.
If we fast forward to today, we see similar debates playing out in economics. When measuring the effects of the minimum wage, some commentators act almost as if measurement is all that matters, as if “Econ 101” supply and demand theory, which would otherwise indicate that artificially high wages mean fewer jobs, can be cast aside. When an empirical study finds no employment effects, they conclude that those who reason based on the theory of supply and demand must be ideologues. But are they? Or are our data and measurement techniques too imperfect to detect the law of downward sloping demand in a fast-paced world?
These kind of issues are also at play in BCA. Open up a textbook and it will likely tell you that the welfare measure underlying BCA is economic efficiency, which relates to maximizing a broad conception of society’s overall wealth. But BCA, at least as it is produced in the government and in countless academic studies, is not measuring efficiency.
This is true for at least two reasons: First, BCA does not properly account for the opportunity cost of capital, detailing how capital would be employed with and without a government policy change. Second, BCA applies weights to consumption based on when the consumption occurs in time.
With respect to the opportunity cost of capital, economists have understood, since the early 1960s, that the proper way to account for the opportunity cost of capital in analysis is by using a shadow price, which is a factor by which all capital benefits and costs are converted to equivalent units of consumption. Sometimes this conversion device is called a marginal cost of funds factor. Whatever the name used, without such a conversion, comparing one dollar of capital to one dollar of consumption is comparing apples to oranges.
The government tries to address this issue by using a discount rate, but its approach is almost always inappropriate. Units of consumption and units of capital are all discounted at the same rate, as if they are growing over time at the same rate. But consumption benefits dissipate quickly, while the returns to capital can increase with time as some are reinvested. Treating these different benefits as if they are the same gives too much weight to consumption—capital increases social wealth by more.
Through discounting, the analyst also applies a set of weights to consumption streams based on who receives them and when. If John receives a consumption benefit worth one dollar today, this benefit receives more weight than if Sally receives an equivalent benefit next year. But is Sally’s consumption really so different from John’s? Even if it is, a government analyst is ill-equipped to distinguish how this experience varies across people.
From the standpoint of economic efficiency, a dollar’s worth of consumption should always receive the same weight (adjusting for inflation, of course). Equal weighting along these lines is standard within a time period but for some reason it becomes controversial across time. Analysts must be careful not to attribute characteristics of individuals—like time preference, diminishing marginal utility, or risk aversion—to society as a whole, an error known as a fallacy of composition. Ironically, Koopmans himself, through his influential work on time preference, likely contributed to this tendency of analysts to blur individual and social characteristics.
BCA has been a fundamental part of the regulatory process in the U.S. federal government since the early 1980s. Executive Order 12,866, which governs the U.S. regulatory analysis and review process has just enjoyed its 25th anniversary. The government’s benefit-cost watchdog, the Office of Information and Regulatory Affairs, has existed for nearly 40 years.
And yet, after all these years, what exactly is BCA measuring? If not efficiency, then what? Without a clear welfare measure, BCA is like a rudderless boat adrift at sea. It can be a useful tool, but to be truly useful in practice, first BCA has to measure something meaningful in theory.
ON THE ECONOMY: DON’T BRING ME DOWN
John Dunham, Managing Partner, John Dunham & Associates
You got me runnin’ goin’ out of my mind. You got me thinkin’ that I’m wastin’ my time. Don’t bring me down, no no no no no. I’ll tell you once more before I get off the floor – Don’t bring me down. These lyrics begin the song – you guessed it – Don’t Bring Me Down, the ninth track on the English rock band the Electric Light Orchestra’s 1979 album Discovery. The song was written by Jeff Lynne and was dedicated to the Skylab space station, which re-entered the Earth’s atmosphere that same year.
One thing that has not entered the atmosphere during the current business cycle is interest rates. Over the course of the 10 year business cycle (from the bottom of the last recession) the benchmark 10-year rate has barely budged. The 10-year yield was 2.18 percent in December 2009, at the very bottom of the recession and it is at just 2.75 percent today. Over the course of the growth portion of the business cycle, the 10-year yield fell to as low as 1.38 percent. With inflation averaging about 2 percent, interest rates over the course of the business cycle have essentially been negative.
While interest rates overall have been falling for years, they do tend to rise over the course of a recovery, as consumers and firms increase borrowing to fund new purchases and investments. But beginning in 1987, things began to change, and during three of the last four recoveries, interest rates continued to fall. What happened in 1987? Well, Ben Bernanke became Chairman of the Federal Reserve Board of Governors, and while Bernanke and the Chairmen who have followed him have all been dovish on monetary policy, they don’t really have a huge impact on market interest rates. Rather the Fed only sets overnight rates, and while these can eventually flow into market rates, it is really the demand for money and inflation that determine rates at the 10-year level.
Looking at the numbers historically, a large part of the downward trend came about through rates climbing down the mountain that occurred in the 1970s. Inflation peaked in 1980 at about 14.5 percent. At that time the 10-year yield approached 15.5 percent, a real interest rate of about 1 percent. Today, inflation sits at roughly 2 percent, and the 10-year yield is 2.75 percent, a real rate of about 0.75 percent. So in effect, nominal interest rates have consistently fallen since 1987 because inflation has consistently fallen over the same period.
This explains the long-term trend, however, since inflation tends to rise as the business cycle matures why not interest rates as well? If demand for money rises, should not prices? And with debt levels skyrocketing over the same period, it is incongruous that the price of money has continued to fall.
The only thing that could lead to the decline over the course of the business cycle is that the supply of capital has outstripped demand. Statistics on the supply of money are a bit iffy since there are many different measures; however, even examining the broadest measure of the money supply, commercial debt alone has grown by 1.5 times the supply of money.
So standard economic theory does not seem to apply to interest rates. Reduced inflation rightfully brought nominal interest rates down; however, demand for debt should bring real interest rates up. Stanley Fischer, former Vice Chairman of the Fed suggested that slower expected growth due to decreased productivity and low labor force participation, is the main cause of low interest rates since investors do not see much demand for money in the future. While this may have some effect, it is unlikely to offset current debt to savings ratios. Fischer also suggested that business was becoming less capital intensive so demand for credit was weakening. This does not seem to fit the current scenario where debt is piling up even if it is not being used for productive investments.
One interesting theory is that as information and intangible assets like software or network effects have become a larger part of the capital input of most goods and services, and since these assets are not properly valued in GDP and other financial calculations, they are in effect hidden savings that are supplying capital to the markets at a faster rate than debt has grown. In effect, the value of a Facebook or an Uber is made up almost entirely in the value of their network effects and this is not measured in the capital account statistics. So if Facebook borrows $100 to attract 100 more users it adds $100 to the debt calculation, but nothing to the asset side of the balance sheet. Maybe it is a coincidence that the internet boom began with the first iterations of what became America Online (or AOL) at about the same time as the general fall in nominal interest rates.
Whatever the case may be, low interest rates now dominate financial markets and are undergirding a huge amount of debt. Any marked increase will cause very large dislocations, so hopefully the markets are rational and some sort of interest rate bubble is not u