Research Analyst
Tim is a research analyst in the Research Department of the Federal = Reserve=20 Bank of Cleveland. He joined the bank in May 2009, and his research = interests=20 are financial markets, monetary theory, and econometrics.
Read=20 full bio
Senior Research Economist
Filippo Occhino is a senior research economist in the Research = Department at=20 the Federal Reserve Bank of Cleveland. His primary areas of interest are = monetary economics and macroeconomics. His recent research has focused = on the=20 interaction between the risk of default in the corporate sector and the = business=20 cycle.
Read=20 full bio
= Economic=20 TrendsEconomic=20 CommentaryPolicy=20 Discussion PapersWorking=20 PapersEconomic=20 Review
Subscribe to=20 e-mail announcementsSubscri= be to=20 Research RSS How=20 to subscribe to RSS TwitterSearch Fed=20 publications
03.24.11
Falling home and financial asset prices have = combined to=20 weaken the average household=E2=80=99s balance sheet, and this has = helped to slow down=20 the current recovery. We examine the role that household balance sheets = have=20 typically played in postwar business cycles and assess their importance = in=20 explaining why some recoveries, including the current one, have been = weaker than=20 others.
The slow pace of the current recovery has been a source of concern = for some=20 time. Whereas real GDP growth after severe recessions has generally been = very=20 strong, that has not been the case during the current recovery, which = followed=20 the worst postwar recession. This time around, it took three years for = real GDP=20 to return to where it was just before the start of the recession.
One factor behind the slow recovery has been the weakness of = household=20 balance sheets. During the financial crisis, the values of real estate = assets=20 and financial assets plunged, lowering household net worth and raising = leverage.=20 To repair their balance sheets, households have been increasing their = saving=20 rate, raising the average from its pre-recession level of around 2 = percent to=20 its current level above 5 percent. This deleveraging process has slowed=20 consumption and, as a result, the recovery.
We examine the role that household balance sheets have played in = postwar=20 business cycles and assess their importance in explaining why some = recoveries,=20 including the current one, have been weaker than others. We begin by = documenting=20 some important differences across past business cycles. We then study = the role=20 played by household balance sheets using a simple, conventional model of = the=20 U.S. economy. We highlight both the direct impact of balance sheets on = economic=20 activity and the indirect mechanism through which balance sheets amplify = and=20 propagate the effects of macroeconomic shocks. We find that balance = sheets have=20 contributed importantly to the dynamics of some business cycles, = especially the=20 more recent ones.
Past business cycles have not followed a unique pattern. In = particular, there=20 have been important differences in the recoveries. Some recoveries, = especially=20 those right after World War II, were strong and rapid. Others, = especially the=20 more recent ones, have been weaker and longer.
To compare the different business cycle patterns, we decompose real = GDP into=20 the sum of the GDP trend and the GDP gap.1 The trend is = determined by=20 long-run factors such as structural productivity, capital accumulation, = and=20 long-run growth in the labor force, while the gap is the cyclical = component that=20 reflects short-run factors. Since we are interested in business cycle=20 properties, we take the trend as given and base our analysis on the = gap.
In the earlier postwar cycles (1948=E2=80=931960), recoveries were = short, rapid, and=20 strong (figure 1). Real GDP rebounded right after the trough of the = recession,=20 growing faster than the trend growth rate and rapidly catching up to the = trend.=20 The path of the GDP gap was roughly symmetrical around the trough of the = recession and looked V-shaped: It first decreased and then rapidly = returned to=20 zero, with the recovery phase lasting approximately as long as the = recession=20 phase. The two cycles in the 1980s looked V-shaped as well.
=20
Source: Authors=E2=80=99 calculations from Bureau of Economic = Analysis data.
Recoveries were weaker during the 1970s (1969 and 1973 cycles) and = they=20 lasted longer=E2=80=94that is, GDP took longer to return to trend. The = path of the GDP=20 gap looked more like a checkmark than like a V, with the recovery phase = lasting=20 longer than the recession phase. The longest and weakest recoveries have = been=20 the three most recent ones (1990, 2001, and 2007 cycles). There has been = no=20 strong GDP rebound or growth, and the recovery phase has lasted years. = We might=20 say the path of the GDP gap was somewhat L-shaped; though perhaps it is = more=20 accurate to say it was erratic and shapeless.
Another way to see differences in business cycle patterns is by = comparing the=20 lengths of recessions with their respective recoveries. Table 1 lists = the=20 NBER-measured lengths of all postwar recessions along with two = alternative=20 measures for the length of the subsequent recoveries. The first measure = is the=20 number of quarters that it took for GDP to return to its trend.
Cycle
Length of recession (quarters)
Length of recovery(quarters)
Duration of = recovery(quarters)
1948:Q4
1953:Q2
1957:Q3
1960:Q2
1969:Q4
1973:Q4
1980:Q1
1981:Q3
1990:Q3
2001:Q1
2007:Q4
Source: NBER; authors=E2=80=99 calculations.
The second, which we label the duration of the recovery, is a more=20 sophisticated measure that takes into account whether most of the GDP = gap was=20 closed earlier or later during the recovery phase.2
For all the V-shaped business cycles, especially the first four in = our=20 sample, the length of the recovery tended not to exceed the length of = the=20 preceding recession. In contrast, the recovery lasts longer than the = recession=20 in the other business cycles, especially those after 1990.
Not only has the pattern of GDP over the cycle tended to change over = time, so=20 has the pattern of unemployment. During the earlier cycles, unemployment = peaked=20 approximately at the trough of the recession and then rapidly decreased = (figure=20 2). Since the 1970s, unemployment has peaked later and later and has = decreased=20 more slowly.
=20
Note: Shaded bars indicate recessions.Source: Bureau of Economic=20 Analysis.
Household balance sheets have been one factor behind these business = cycle=20 differences. Balance sheets have been weaker in the slower = recoveries.
To measure the weakness of balance sheets, we use leverage, defined = as the=20 ratio of assets to net worth. High leverage indicates weak balance = sheets.=20 Leverage in the household sector was especially high relative to trend = during=20 the 1973, 2001, and 2007 cycles, whose recoveries were especially slow = (see=20 figure 3). During all of these three business cycles, financial asset = prices=20 experienced sizeable drops. In addition, house prices plunged during the = last=20 cycle. As asset prices fell, the value of households=E2=80=99 assets = decreased, raising=20 households=E2=80=99 leverage and weakening their balance sheets (a = decrease in asset=20 values causes a larger percentage drop in net worth, raising = leverage).
=20
Notes: Leverage is defined as the ratio of assets to net worth. = Leverage has=20 been logged and detrended. The trend has been computed using the HP = filter.=20 Shaded bars indicate recessions.Sources: Board of Governors of the = Federal=20 Reserve System, Flow of Funds; NBER.
When their leverage is high, households increase their savings = because they=20 want to repair their balance sheets. This switch to saving tends to = decrease=20 consumption and to slow down the recovery. In fact, data show that = leverage=20 relative to trend is negatively correlated with present and future = production,=20 so high leverage tends to be associated with and followed by a decrease = in real=20 activity. This type of evidence is not sufficient by itself to identify = the role=20 played by household balance sheets in business cycles, however. It can = show that=20 high leverage is associated with low economic activity, but it cannot = discern=20 whether leverage actively causes changes in economic activity or whether = it=20 passively responds to economic activity.
To formally assess the relationship between balance sheets and the = business=20 cycle, we use a common, simple model of the aggregate economy. The = variables=20 included are production, the unemployment rate, inflation, the = short-term=20 risk-free interest rate, an oil-price index, and household leverage. The = model,=20 a vector auto-regression (VAR), relates each variable to the past values = of all=20 variables in the model and to various types of shocks.3 This type of = model has been=20 shown to effectively capture the dynamics of the business = cycle=E2=80=94we use it to=20 determine how changes in balance sheets affect economic activity and how = balance=20 sheets respond to changes in the economy.
Read Full Article »