The 1990-91 Recession: Gulf War, S&L Crisis, and the First Jobless Recovery

The 1990-91 US recession was the first NBER-dated recession in nearly eight years, ending the long Reagan-era expansion that had begun at the November 1982 trough. It ran 8 months from July 1990 to March 1991, contracted real GDP by 1.4 percent peak to trough, and pushed unemployment from 5.2 percent at the recession start to a peak of 7.8 percent reached in June 1992, fifteen months after the official recession trough. The cycle is often called the Gulf War recession because the August 1990 Iraqi invasion of Kuwait coincided with its start, but NBER identified the peak as July 1990, one month before the invasion, indicating the cycle had begun before the oil shock amplified it.

Three reinforcing factors caused the recession. The savings and loan crisis had stressed the financial system through the late 1980s, requiring federal resolution efforts that absorbed taxpayer resources and tightened banking-sector credit availability. Federal Reserve tightening through 1988-89 (the federal funds rate rose from 6.5 to 9.75 percent) had cooled the economy in the way that postwar Fed tightening cycles routinely precede recessions. And the August 1990 oil-price doubling that followed the Iraqi invasion sharpened the contraction by hitting consumer confidence harder than any single news event since the Cuban Missile Crisis.

The cycle is also notable for being the first jobless recovery in postwar US data. GDP growth resumed in Q2 1991 but employment did not return to its July 1990 peak until February 1993. The phenomenon would be replicated in the 2001 and 2007-09 cycles, suggesting a structural change in how US labour markets respond to business-cycle troughs.

Origins: The Savings and Loan Crisis

The savings and loan crisis was the failure of approximately 1,043 of 3,234 US savings and loan associations between 1986 and 1995. The underlying causes had been building for two decades. S&Ls had historically borrowed at short-term deposit rates (typically 5 to 6 percent in the 1960s and 1970s) and lent at long-term fixed-rate mortgages (typically 7 to 9 percent). The arrangement worked as long as deposit rates stayed below mortgage portfolio yields. The 1981-82 Volcker recession pushed short-term interest rates above 15 percent while existing S&L mortgage portfolios still yielded the historical 8 to 9 percent, producing massive negative carry that wiped out S&L profitability and, for many, capital.

Congress responded with the Garn-St Germain Depository Institutions Act of 1982, which allowed S&Ls to enter commercial real estate, business, and consumer lending in an attempt to rebuild their earnings. The deregulation produced enormous risk-taking by thrifts that had no expertise in commercial lending. Many of the resulting commercial real estate loans (particularly in Texas, Arizona, California, and Florida) defaulted in the late 1980s as oil prices fell and overbuilt commercial real estate markets corrected. Fraud at some institutions (Charles Keating's Lincoln Savings being the most notorious) compounded the structural losses.

By 1989, the Federal Savings and Loan Insurance Corporation was itself insolvent. The Financial Institutions Reform, Recovery, and Enforcement Act of August 1989 created the Resolution Trust Corporation to liquidate failed thrifts. Final taxpayer cost approached $124 billion. The cleanup tightened credit conditions across the broader banking sector, with commercial real estate lending particularly affected. The financial-system stress was the slow-burn pre-condition that the 1990 oil shock then amplified into recession.

The Greenspan Tightening

Alan Greenspan became Federal Reserve Chairman in August 1987, succeeding Paul Volcker. Inflation accelerated through 1988, with consumer price inflation reaching 4.4 percent in early 1989. Greenspan's Federal Reserve raised the federal funds rate from 6.5 percent in early 1988 to 9.75 percent by February 1989, a tightening cycle designed to slow inflation before it could reach 1970s levels. The tightening had three intended effects: it slowed credit growth, pressured the already-stressed thrift sector, and cooled the commercial real estate market that had been driving late-1980s capital expenditure growth.

By mid-1990, the Fed had begun cutting modestly (the funds rate was at 8.0 percent in July 1990). But monetary policy operated with the usual 12 to 18 month lag, and the late-1980s tightening was still being absorbed by the economy. Housing starts had been falling for two years. Commercial construction had pulled back. Consumer credit growth had slowed. The economic conditions that the recession dating committee subsequently identified as the cycle peak were in place by early 1990.

The Iraqi Invasion and the Oil Shock

On 2 August 1990, Iraqi forces invaded Kuwait. The invasion immediately removed Kuwaiti oil production from world supply and threatened Saudi oil production via the perceived Iraqi threat to Saudi territory. Crude oil prices rose from approximately $17 per barrel in mid-July 1990 to a peak above $36 per barrel in October 1990, a 112 percent increase in three months. The price shock was the largest since the 1979-80 Iranian Revolution oil shock and the third major oil-supply disruption in seventeen years.

Consumer confidence fell from a Conference Board reading of 100 in July 1990 to 61 in October 1990, the sharpest two-month decline since the Cuban Missile Crisis of 1962. Household spending on durable goods (autos, appliances, furniture) pulled back sharply as households absorbed higher gasoline and heating costs and braced for what many anticipated would be a wider Middle East conflict. Business capital expenditure slowed as firms postponed capacity expansion in the face of demand uncertainty. The combination of pre-existing financial-system stress and the oil-shock-driven confidence collapse produced the contraction that NBER subsequently identified as beginning in July 1990.

The Gulf War and the Recession Trough

Operation Desert Storm began on 16 January 1991 with US-led coalition airstrikes against Iraqi forces. The ground war began on 24 February and Kuwait was liberated by 27 February. The war was over within six weeks, dramatically shorter and less costly (in terms of US casualties and oil-infrastructure damage) than most forecasters had anticipated. Oil prices fell from their October peak above $36 per barrel back to approximately $20 per barrel by February 1991. Consumer confidence rebounded from its November trough. The Federal Reserve had been easing aggressively (the funds rate fell from 8.0 percent in mid-1990 to 6.0 percent by early 1991, on its way to 4.0 percent by 1992).

By March 1991, the recession reached its NBER-dated trough. Real GDP had contracted 1.4 percent peak to trough, modest by postwar standards. The contraction was concentrated in interest-rate-sensitive sectors (housing, autos, commercial construction) and in the regions most exposed to commercial real estate stress (Texas, Arizona, the Northeast). Manufacturing employment fell less sharply than in earlier recessions reflecting the secular shift toward services.

The Jobless Recovery

The recovery from the March 1991 trough was unusually slow on the labour market. GDP growth resumed in Q2 1991 and continued at modest rates through 1992. But nonfarm payrolls did not return to their July 1990 peak until February 1993, nearly two years after the recession officially ended. Unemployment continued rising after the trough, peaking at 7.8 percent in June 1992, fifteen months later. The phenomenon was unprecedented in postwar US data: every previous postwar recession had seen employment recover within roughly twelve months of the GDP trough.

The leading explanations for the 1990-91 jobless recovery are several. First, the lingering financial-system stress (S&L resolution, commercial real estate workout, bank capital rebuilding) constrained credit availability for new hiring and expansion through 1992. Second, structural adjustments in manufacturing accelerated during the cycle: the shift to flexible manufacturing, automation, and offshoring of some production processes reduced labour-intensity per unit of output. Third, white-collar middle management layoffs of the late 1980s and early 1990s (a new phenomenon at the scale that occurred) represented permanent reductions rather than cyclical adjustments. The labour-market weakness contributed materially to George H W Bush losing the 1992 presidential election despite the recession officially ending in March 1991, with Bill Clinton's campaign focused on the still-weak labour market.

Lessons from the 1990-91 Cycle

Three lessons from the 1990-91 cycle still inform business-cycle thinking. First, financial-system stress operates with long lags and can shape macroeconomic outcomes years after the underlying problems emerge. The S&L losses began accumulating in 1981-82; the resolution dragged through 1995. The macroeconomic consequences played out across more than a decade. The pattern repeated in the 2007-09 cycle on a much larger scale, with mortgage-credit losses that began in 2006 producing macroeconomic consequences through 2015.

Second, the jobless recovery phenomenon that first appeared in 1990-91 has become a structural feature of US business-cycle dynamics. The 2001 and 2007-09 cycles replicated it more severely. The 2020 COVID recovery appeared to break the pattern on aggregate employment numbers, but the sectoral and demographic divergences (the K-shaped recovery) revealed labour-market heterogeneity that mirrored earlier jobless-recovery dynamics.

Third, geopolitical events tend to compress or amplify recessions that are already underway rather than cause recessions from scratch. The 1990-91 cycle had begun in July 1990 before the Iraqi invasion. The 2001 cycle had begun in March 2001 before 9/11. The 2020 cycle was directly caused by the COVID public-health event, the exception to the pattern. The general principle is that supply shocks operate on a baseline of pre-existing macroeconomic conditions, and the conditions matter as much as the shock for the resulting business-cycle outcome.

For comparison with other post-WWII cycles, see post-WWII recessions overview. For the next major cycle, see the 2001 recession. For the deepest postwar contraction, see the 2007-09 Great Recession.