FIRST EXPERIENCES
My career in forecasting over the past six decades has roughly coincided with its increasing prominence in both the private and public spheres. I was one of the founding members of the National Association for Business Economics in 1959, and I became its president in 1970. I also became chairman of the Conference of Business Economists in 1974 and was scheduled to be chairman of the Economic Club of New York in 1987 had I not joined the Federal Reserve. While business economists in those early years dabbled in macroforecasting, most of us were much more heavily focused on microforecasting—for firms and for industries. Macroforecasting remained largely in the hands of academia and the government.
MACROFORECASTING
My first experience with macroforecasting was more confusing than edifying. As World War II was winding down, Keynesian macroforecasters predicted that with the end of military spending, a “mature” America, as the eminent Keynesian Harvard economist Alvin Hansen put it, would slump back into the “secular stagnation” that prevailed in the prewar world of the Great Depression.8 Hansen’s thesis was widely supported and I, as a college student, found his arguments persuasive. My first tutorial in contrary opinion, however, arose when I read an equally convincing tome by a nonacademic microeconomic practitioner, George Terborgh of the Machinery and Allied Products Institute. His book The Bogey of Economic Maturity was published just after the war. In the event, the U.S. economy boomed in the postwar period. The origins of that boom were probably more nuanced than Terborgh had argued and less damning to Hansen’s view of the economy than might appear on the surface. But the lesson that I drew was that when it came to macroeconomic forecasts, intelligent people could make big mistakes.
TO WORK
My first job out of college in 1948 was with the highly respected business research organization known today as the Conference Board. I left the Conference Board in 1953 and joined William W. Townsend, a veteran Wall Streeter forty-one years my senior, to form a small consulting firm, Townsend-Greenspan & Company. We had a wonderful relationship for five years until he passed away in 1958. I carried on, little knowing what lay ahead. I was finally in the forecasting business, though not yet as a macroforecaster. However, my forecasting experience was widening at the industry level.
I look back fondly to the 1950s and 1960s, when I specialized in individual industry (micro) forecasting with a bit of finance thrown in. I immensely enjoyed delving into how individual markets worked at a level of detail not feasible in macromodels. In my early years, when steel was a specialty, I read all eight hundred pages of the industry bible, The Making, Shaping and Treating of Steel. Years later, in 1997, when appearing before a meeting of the American Iron and Steel Institute, I asserted with no small amount of nostalgia that I was the only Federal Reserve chairman who had read that book from cover to cover. Nobody challenged my assertion.
As the years rolled on, Townsend-Greenspan became quite diverse, employing micromodels to help analyze a wide variety of markets—some global—in oil, natural gas, coal, pharmaceuticals, and motor vehicles, though not yet high tech, whose prominence was still in the future. We were increasingly successful through the 1960s dispensing our brand of microeconomics, and we prospered.
The broad scope of industries we eventually covered inevitably brought us close to forecasting macrodevelopments. But using the same techniques we devised to evaluate individual industries did not always go well when expanded to the global stage. We had, for example, assumed that with global demand for petroleum products being demonstrably historically price inelastic (that is, unresponsive to price change), the rapid escalation of oil prices that emerged from the OPEC oil embargo of 1973‒74 could continue for years. But oil demand, to my surprise, fell off sharply as prices rose, indicating that the demand for oil was more elastic with respect to price than I and many others had previously appreciated. The rate of oil consumption per dollar of real GDP unexpectedly turned downward, lessening pressures on inflation. I was well wide of the mark on that forecast.
COMPETITORS
Our forecasting competitors also made some massive errors from time to time. In retrospect, some of our more important collective errors reflected an incomplete understanding of the growing complexity of finance. Even the economists for banks and financial institutions did not distinguish themselves as great forecasters of either the economy or financial markets. The two presumably more sophisticated computer-based macroeconomic forecasting firms—Data Resources Inc. (commonly known as DRI) and Wharton Econometric Forecasting Associates (both founded in 1969)—had their share of forecast misses but nonetheless built formidable successful and econometric-driven firms.
DRI and Wharton Econometrics remained at the forefront of the macroeconomic forecasting business for decades. DRI was headed by Otto Eckstein, a noted Harvard professor and member of President Johnson’s Council of Economic Advisers. Wharton Econometrics’ roots were more academic than those of its competitor. Founded by Lawrence Klein9 of the Wharton School, the firm grew out of the school’s economics research unit. That unit had received so much sponsorship from U.S. corporations in the preceding few years that it needed a more defined structure to manage the burgeoning volume of projects coming through its doors. DRI grew until eventually merging with Wharton Econometrics in 2001.10
As DRI and Wharton continued to grow, so did Townsend-Greenspan and our microforecasting work.
THE FORD YEARS11
I carried my statistical experiences with me as I took a leave of absence from Townsend-Greenspan to become, in September 1974, chairman of President Gerald R. Ford’s Council of Economic Advisers. Shortly thereafter, the economy was in trouble. New orders received by businesses for their products were declining, production began to fall rapidly, and unemployment started to increase in discontinuous jumps. That the economy was heading into a recession (if it were not in fact already in one) didn’t require much debate.
As 1974 drew to a close, retail sales and home building were soft, and much of what we consider final demand was slipping. By Christmas 1974, the key question for economic policy was whether we were experiencing an inventory recession, which meant a sharp but temporary erosion in production and employment as businesses worked off excess inventories, or a far more dangerous softening in the economy brought about by a more persistent weakness in final demand. This was the burning issue for President Ford. An answer had to be formulated as quickly as feasible because the types of economic policy initiatives that one should employ depended on the answer.
For a short-term inventory recession, the optimum policy, as we saw it, was to do as little as politically possible and let the natural forces of the economy bring the recession to a halt. If it looked as though the bottom were falling out of final demand, much more drastic policy options would have to be considered. The problem was that I felt our existing economic intelligence was inadequate to monitor the rapidly weakening economy.
The political advice being offered to the administration was unequivocal. George Meany, the president of the AFL-CIO, was typical. “America is in the worst economic emergency since the Great Depression,” he testified in March 1975. “The situation is frightening now and it is growing more ominous by the day. This is not just another recession, for it has no parallel in the five recessions in the post‒World War II period. America is far beyond the point where the situation can correct itself. Massive government action is needed.”
The Council of Economic Advisers initially didn’t have even a monthly GNP series to guide policy, but starting in December 1974, we developed what amounted to a weekly GNP. It may not have passed the rigid statistical standards of the Bureau of Economic Analysis (BEA) of the Department of Commerce, but it was more than adequate—in fact quite instrumental—in answering the question of whether we had an inventory recession, a final demand recession, or both.
While the Department of Commerce has since abandoned its weekly retail sales series, it nonetheless did yeomanlike serv
ice during that period in indicating that personal consumption expenditures were not undergoing a downward plunge. The other sectors of the economy had to be estimated more indirectly. Industry trade sources, coupled with the latest data on building permits and housing starts, outlined the residential sector for us on a weekly basis. Survey forecasts of plant and equipment, monthly new orders and shipments for machinery, data on nonresidential construction, and, with a delay, imports of capital equipment were a crude proxy for capital investment. From the unemployment insurance system we were able to get a rough indicator of aggregate hours worked, which, combined with an estimate of output per work hour (which were little more than educated guesses), yielded a rough estimate of total real GNP, which was then reconciled with its component parts.
Putting all of these statistics together indicated, with some degree of robustness, something that we knew for a fact only much later: that the rate of inventory liquidation—the gap between GNP and final demand—was exceptionally large by historic standards. That gap reflected the fact that production had been cut well below the level of final demand in order to work off the excess inventories that had accumulated. Therefore, if final demand continued to stabilize, as apparently it was doing in the early weeks of 1975, the recession’s low point was close at hand and a marked rebound from the downturn was highly likely. Inventory liquidation cannot go on indefinitely. It must eventually slow, and that process closes the gap between final demand and production. It soon became clear from the weekly insured unemployment data and several qualitative indicators that the worst was over.
At that point we could conclude that further expansionary measures would be unnecessary and in the long run could turn out to be counterproductive.12 Short-term emergency GNP monitoring was no longer necessary, and the short history of the weekly GNP came to a very creditable end.13
POLITICS
Looking back on that period, President Ford exhibited unusual political courage in 1975 by acquiescing in only a very small stimulus when “conventional wisdom” at the time was for a policy response that was far more aggressive. Ford nonetheless persevered. In much the same way, so did Ronald Reagan by his unwavering support of Paul Volcker and the Federal Reserve’s tight money policy of 1980‒82, when political opposition to the Fed’s efforts was most intense.
In any democratic society, it is very difficult for presidents, congressmen, central bankers, or any other economic policy makers to move ahead of conventional wisdom, which is often reinforced by the pressures of herd behavior. To lead markets requires a conviction that is rare among public officials because running ahead of markets, of necessity, implies holding a view contrary to the participants in markets themselves, who invest real money. And in my experience, if policy makers are in a minority and wrong, they are politically pilloried. If they are in a majority, and wrong, they are tolerated and the political consequences are far less dire. Presidents Ford and Reagan were more politically courageous than I had realized at the time.
When in public office myself, I became acutely aware of the pull of political bias. Politicians are driven to take a stance that (1) the outlook is unequivocally good (the result of their policies) and no change in policy stance is required (an upbeat assumption that is always acceptable, even if proved wrong), or that (2) things are not good and nothing short of maximum response is acceptable, for if events do in fact turn dire, half measures appear as (politically) ineffectual as no action at all.
Most nonelected public policy makers, in my experience, are able to fight this political bias but probably never fully successfully. I tried to insulate myself when I became chairman of the Council of Economic Advisers in 1974. I told Ford’s chief of staff, Donald Rumsfeld, that I could not continue the practice of the CEA chairman’s being the administration’s chief economic spokesman, because I was certain that I would differ with, and hence could not support, some of the president’s economic initiatives—the WIN (Whip Inflation Now) buttons, for example. As I requested, the task was given back to the secretary of the treasury. But much to my surprise and delight, there were very few economic initiatives pursued by President Ford with which I strongly differed.
BACK TO PRIVATE ENTERPRISE
Ford’s term as president ended at noon on January 20, 1977, and I was on the noon shuttle to New York that day and by 2 p.m. was back in my old office at Townsend-Greenspan overlooking New York Bay. I immediately fell into my pregovernment routine of following and forecasting the American—and later the world—economy. In fact, my daily activities were not that much different from what I had pursued in Washington for two and a half years. The sole innovation was that I was devoting time to construct a macroeconometric model and created a computer-based forecasting system for the first time. While Townsend-Greenspan remained active in industry-based microforecasting, my efforts to develop a macroeconometric model grew out of my growing interest in understanding the workings of the economy as a whole and the complementary benefits that would flow to our industry-based work.
I joined the boards of Mobil, JPMorgan, Alcoa, General Foods, Capital Cities-ABC, and ADP. Most were, or became, clients of Townsend-Greenspan. After President Reagan was elected, I was appointed a member of his Foreign Intelligence Advisory Board, much of the time advising on the accuracy of foreign statistics, especially those of the Soviet Union. In addition, during the Reagan administration years I served on numerous presidential commissions, most important, the Social Security Reform Commission (1983).
AT THE FED
I joined the Federal Reserve in August 1987, two months before the bottom fell out of the stock market. The stock market decline of more than one fifth on October 19, 1987, created a massive contraction in market wealth. Accordingly, the Fed opened up the money spigots to full throttle. I was deeply puzzled as the weeks went on by the small effect the capital losses had on economic activity. It was not until I experienced the demise of the dot-com and housing bubbles that I finally concluded that the reason for the economy’s unresponsiveness to the shock of October 19, 1987, was that the stock market investors who absorbed the huge losses at that time were wholly unleveraged.14
On becoming chairman of the Federal Reserve Board, I soon realized that it had as an integral part of its organization something like 250 of the most skilled PhDs in economics, certainly in the United States and very probably in the world. That group was led through my eighteen-year tenure by Mike Prell and David Stockton as the heads of the Division of Research and Statistics, and Ted Truman and Karen Johnson as heads of the board’s International Division. There wasn’t a question I could imagine to which I couldn’t find some expert to respond somewhere in the organization. Domestic forecasting was directly or indirectly processed through the Federal Reserve’s macroeconomic model of the U.S. economy, though there was extensive tweaking of the model’s output to capture economic forces not captured by the model.
Despite the fact that the model missed the collapse of 2008 along with virtually all other models, its historical record has been better than most.
While decisions of the Federal Reserve’s Federal Open Market Committee (FOMC) are not legally subject to alteration by any other agency of government, it has always been clear that lurking in the background is the possibility that a Congress and/or a president, irate at monetary policy, could, through legislation, alter the Fed’s degree of independence, thereby crippling its effectiveness. In my eighteen and a half years as Fed chairman, I received a figurative truckload of requests from Congress urging an easier monetary policy. I don’t recall ever receiving a single request urging the Fed to tighten.15
The Fed grew concerned in August 1991 when Senator Paul Sarbanes introduced a bill to restrict FOMC votes to governors only. This would be an irreversible step toward chronically easier monetary policy because, in my experience, the evidence does support the view that presidents of the Federal Reserve banks have been more “hawkish” than the presidentially appointed (and Senate confirmed) governors. The
bill did not muster the support needed to become law. With the economy expanding, at times briskly, threats to the Fed’s independence from mid-1991 to 2008 were minimal.
But, as I noted in the Introduction, the Fed came under increasing pressure for greater congressional oversight when, in response to the financial breakdown of 2008, it invoked the rarely used provision of the Federal Reserve Act that enables it to lend virtually without limit to virtually anyone, in the United States or abroad. The Fed lent $29 billion to JPMorgan to facilitate the bailout of Bear Stearns in March 2008. I viewed it as the Fed’s acting in its long-honored role as fiscal agent of the U.S. Treasury, and would have preferred the U.S. Treasury to immediately swap Treasury securities with the Fed for its claims against JPMorgan.
When I raised the issue of the Treasury’s quickly taking over the central bank’s commitments, I was told by a senior Treasury official that that would have required the administration to request appropriated funds from the Congress (true)—a political nonstarter. That was a most regrettable failure, especially because it was not a matter of substance but a quirk in fiscal bookkeeping. Whether claims are held as an asset of the central bank or by a subsidiary of the U.S. Treasury makes no difference to markets or taxpayers. The liabilities of the Treasury and the Federal Reserve are both interchangeable sovereign liabilities of the U.S. government.
In the aftermath of the financial crisis, much of the Federal Reserve’s independence appeared for a time to be at risk of being severely weakened. In the event, however, the response of the Congress with respect to the Federal Reserve has been fairly muted to date. The basic structure and functioning of the system has remained intact, including the voting power of the bank presidents.
The Map and the Territory Page 12