- Growth in physical capital per worker has contributed the most to U.S. productivity growth.
- U.S. capital accumulation is increasingly dependent on foreign capital inflows.
- If future technology improvement occurs at its average historical rate, maintaining U.S. productivity growth will require more rapid capital accumulation, especially because worker efficiency appears likely to stagnate or decline.
Not Enough Shovels? The Crucial Role of Future Capital Accumulation for U.S.Output Growth
Productive firms combine physical and human capital with technological knowledge to produce valuable consumer goods and services from apples to Apple smartphones. Producing these goods and services requires machines (physical capital) and skilled workers (human capital). Physical capital includes manmade and natural resources, for example, land, buildings, and machines. Human capital refers to the collection of skills that workers possess and apply to the production process.
Physical capital boosts worker productivity: two workers equipped with two shovels can dig a hole faster than sharing one shovel. History shows this to be valid economy-wide: U.S. output growth has outpaced labor force growth mainly because of increased capital intensity, that is, physical capital per worker. Increased capital intensity was the largest contributor to productivity growth followed by improved technology and labor efficiency. Indeed, the two latter contributors – technological innovations and worker skills – are themselves dependent on prior capital accumulation.
Capital accumulation, in turn, requires prior saving. But U.S. saving and investment are declining precisely at a time when faster capital accumulation is needed to maintain growth in national output and living standards. The baby boomers are on the brink of retirement, and human longevity is improving rapidly. A growing share of retirees in the population will promote higher consumption and reduced saving. Meanwhile, the share of young active workers in the population and their labor efficiency are projected to slow or decline and impose a drag on national output growth. Only by continued capital inflows from foreign sources can the United States maintain growth in the economy and living standards. In general, although growth in human capital is obviously important, advancing physical capital accumulation appears to be critical for maintaining U.S. economic growth.
Saving for the Future
Most people wish to ensure adequate resources for future consumption, especially during old age when they are unable to work and earn. According to one prominent study, this “life-cycle” motive accounts for almost all of the capital accumulated in our economy.1 Only a small fraction of that capital arises because of bequests, as some people die before they have consumed their life savings.
Many goods are perishable, most depreciate over time, and services are almost never storable. Hence, saving for retirement requires purchasing financial claims on future output. Although purchasing and holding financial assets over a long time horizon is risky, the risks are normally smaller in developed nations with better-regulated financial markets. This is partly due to the wide variety of financial securities available to savers in these countries – such as government insured bank deposits, shares and bonds of private firms, government bonds, annuities and whole-life insurance policies, and employer-sponsored defined benefit and contributory pension plans that enable better diversification of retirement savings.
Additionally, recognizing the risks involved in saving for retirement, governments in developed countries have enacted safety-net systems to provide retirement, health care, and other benefits for protecting older generations’ living standards. The value of such entitlement programs in preventing poverty among older generations becomes immediately apparent during severe economic downturns. However, their negative economic effects, which accumulate gradually but inexorably in terms of reduced labor supply and smaller saving and capital formation, are not widely appreciated.
Saving in the United States
Increasing longevity is a good thing, but it also requires people to make smart choices when saving to support consumption during life’s golden years. One must decide how much to save based on myriad factors such as career length, expected health and longevity, interest rates, potential investment returns, inflation, and the risk of capital losses – all of which affect the need and attractiveness of saving for the future. A larger expected return for a given risk level usually elicits more saving.
Typically at the economy-wide level, returns to saving (interest rates and other market returns) adjust to equilibrate the supply of saving with investment demand, rising when investor demand for funds is high relative to the supply of saving and falling when saving is high in relation to investment demand. Today’s high global saving, primarily by residents of emerging economies, has depressed global interest rates to very low levels and reduced that inducement for American residents’ to save more.
Figure 1: Private and Government Consumption Expenditures and the U.S. Net National Saving Rate
Source: PWBM calculations based on information from the Bureau of Economic Analysis.
The first panel of Figure 1 shows that the U.S. net national saving rate (NNSR) has been on a downward trajectory since the mid-1960s.2 And the second panel of Figure 1 shows that government consumption expenditures have declined gradually as a share of net national product (NNP) while private consumption expenditures have increased quite rapidly.
Why has U.S. private consumption increased by so much and for so long? One reason is that the federal government has consistently expanded existing debt-funded social protection programs and enacted new ones during the latter half of the twentieth century (the postwar period). Prime examples are Social Security and health care benefit expansions – all of which have increased the generosity of benefits awarded to older Americans.3 Because a large part of private saving is motivated by “life-cycle” goals – that is, for financing consumption and health care expenditures during retirement – sizable government provision of retirement support may naturally be expected to reduce private saving for those purposes. Indeed, studies of why private consumption has increased suggest a major role for government fiscal policies that promise increasingly generous retirement and health care benefits especially to older Americans.4
National saving would not decline if promises of generous government retirement and health care benefits are funded through corresponding increases in government saving. However, U.S. federal entitlement promises are mostly unfunded, promising to pay benefits to retirees without setting aside commensurate funds to pay them. Indeed, the federal budget has consistently registered deficits during the postwar period and federal debt has skyrocketed.5
While taxes may be increased in the future to pay promised entitlement benefits, taxes today are quite low because the government has also enacted several tax cuts and exclusions from taxable income during the postwar period. Tax cuts were most often motivated by the perception that government action was necessary to help the economy to recover from economic recessions by spurring private consumer spending. Periodic expansions of exclusions of income subject to taxes during the post war period, which benefit all taxpayers regardless of age, have also boosted private consumption spending as a share of NNP.6
The Congressional Budget Office’s latest long-range projections show federal deficits increasing steadily and pushing federal debt above 100 percent of NNP by 2040.7 When unfunded entitlement obligations are also included, U.S. federal indebtedness amounts to several multiples of NNP.
The opposite appears to be true in emerging nations such as China and India where low public provision of social protection programs and relatively smaller income tax exclusions are associated with very high private saving rates. Fortunately for capital formation in the United States, a large share of foreign saving is directed toward financial instruments issued by American financial intermediaries, private firms, and the U.S. Treasury.
Capital Accumulation in the United States
Figure 2: United States’ Net National Saving, Net Domestic Investment and Foreign Funded Component of Net Domestic Investment – Rates Out of Net National Product
Source: PWBM calculations based on information from the Bureau of Economic Analysis.
Low saving appears to be constraining U.S. domestic capital accumulation. Simply put, capital accumulation involves two steps. First, someone must save by consuming less than current income. That saving, usually in the form of financial capital (money) is transferred by banks and other financial intermediaries to domestic investors who use it to purchase or build physical capital such as buildings, machines, and computer software. New and potentially more efficient physical capital can then enhance worker efficiency in producing consumer goods and services. Therefore capital accumulation cannot occur unless someone saves. But U.S. domestic capital accumulation can exceed Americans’ saving if foreigners channel their saving to dollar denominated financial assets – those issued by American investors.
The first panel of Figure 2 shows that the U.S. Net Domestic Investment Rate (NDIR) has exceeded the U.S. NNSR since the early 1980s. The second panel of Figure 2 shows that the Net Foreign Investment Rate (NFIR) – the difference between NDIR and NNSR – has trended upward since the mid-1960s.
It is interesting to note that only a part of incoming foreign saving is channeled into U.S. domestic investment. During 2014, for example, calculations based on National Income and Product Account information from the Bureau of Economic Analysis suggest that of the $401 billion foreign capital inflow, $208.0 billion was invested. The rest of 2014 foreign capital inflows helped to fund American consumer spending during the same year.
Capital Intensity and Output Growth
While additions to the U.S. capital stock have slowed, the existing capital stock has contributed significantly to U.S. productivity growth. The key productive resources are worker hours, capital services, and the technology embodied in those two inputs. But in order to gauge the potential for growth in living standards, it is more useful to focus on labor productivity growth, such as how rapidly output per hour worked is increasing.
Labor productivity (measured as NNP/labor hours) increases when labor efficiency (worker skills) improves, capital intensity (physical capital available per worker) increases, or the technology embodied in those two inputs is better. For capital intensity to increase, capital accumulation must progress at a faster pace than growth in worker hours. This has indeed been the case in the United States, as the first panel of Figure 3 shows. The U.S. capital-labor ratio has risen at an annual rate of 2.0 percent per year.8 Figure 3 also shows that output per hour (labor productivity) has increased almost in tandem with capital intensity.
Figure 3: Capital-Labor and Output-Labor Ratios in the U.S. Business Sector and Contribution of Capital Intensity to Output Growth
Source: PWBM calculations based on data from Bureau of Labor Statistics, June 2015.
How much has growth in physical capital contributed to U.S. output growth? First, note that annual output growth averaged 2.65 percent during 1988-2013, while labor hours grew at 0.85 percent per year.9 This implies labor productivity growth of 1.80 percent per year.10
The second panel of Figure 3 illustrates the contributions of the three productive factors to labor productivity growth: labor efficiency, greater capital intensity, and technology improvements. Of the 1.80 percent growth in labor productivity during 1988-2013, the contribution of labor efficiency improvements averaged 0.29 percentage points; that of technology advances averaged 0.72 percentage points; while the contribution of greater physical capital intensity averaged 0.79 percentage points.
The second panel of Figure 3 also clarifies that increased capital intensity contributed significantly to the “new-economy” after the mid-1990s and was followed by rapid technological advances during the mid-2000s. An emerging concern is that growth of capital intensity has slowed (first panel of Figure 3) and capital’s contribution to U.S. output growth has declined (second panel of Figure 3) after 2009.11
During the near term, growth in the size of the workforce will depend upon immigration and the rate at which boomers retire. In the long term, however, fertility, immigration, population health, and trends in the age of retirement will determine workforce growth. Weakening fertility and limits on the rate of immigration since the turn of the century make slower growth in the U.S. workforce more likely. Projections of the population’s structure – fewer experienced workers after the boomers retire – and of the composition of the population – especially, family structure and labor force participation — also point to slower improvements in labor efficiency. Even if technology improvements are maintained in the future, these labor market projections imply greater dependence on increasing capital intensity to maintain economic growth in the United States.
Projections of U.S. domestic capital accumulation, however, indicate continued slow growth driven by changes in the population's composition. Baby boomers approaching retirement have already accumulated their retirement nest eggs. They will begin to draw down savings during retirement reducing capital formation.12 Other demographic changes also appear consistent with slower domestic capital accumulation: faster growth in population groups with low life cycle saving propensities. If these projections are realized, dependency on foreign capital inflows for maintaining employment rates and output growth will continue to increase in this country. Reducing government deficits and pre-funding future retirement bligations would likely increase the level of domestically supplied capital.
John Karl Scholz, Ananth Seshadri, and Surachai Khitatrakun, “Are Americans saving optimally for retirement?” Journal of Political Economy 114, no. 4 (2006): 607-43, available at: https://www.ssc.wisc.edu/~scholz/Research/Optimality.pdf ↩
NNSR is calculated by first subtracting private and government consumption expenditures from Net National Product (NNP) and dividing the residual by NNP. While the Gross Domestic Product (GDP) is the most commonly used metric in measuring output produced within the territory of a nation, the Net National Product (NNP) reflects the output produced and owned by its citizens, both within the nation’s boundaries and overseas. In addition, NNP excludes capital depreciation and thus reflects the amount of output available to be consumed or saved by the nation’s citizens. Government consumption includes consumption of public goods and services through federal, state, and local government expenditures. ↩
Medicare (1960s), Social Security (1970s), Supplemental Security Income (1970s), Disability Insurance liberalization (1980s), Prescription Drug Benefit (2000s), and more recently the Affordable Care Act (2010). ↩
Jagadeesh Gokhale, Laurence J. Kotlikoff, and John Sabelhaus, “Understanding the Postwar Decline in U.S. Saving: A Cohort Analysis,” Brookings Papers on Economic Activity, no. 1 (1996), available at: https://www.brookings.edu/wp-content/uploads/1996/01/1996a_bpea_gokhale_kotlikoff_sabelhaus_bosworth_haveman.pdf ↩
Since 1960, small federal budget surpluses were accrued during just 5 years: 1969 and 1998-2001. See Table B-19 in the Economic Report of the President, 2015: https://obamawhitehouse.archives.gov/sites/default/files/docs/cea_2015_erp_complete.pdf ↩
Although several tax increases were also enacted, the net effect since the 1960s has been to reduce federal non-social insurance receipts as a share of NNP. Prime examples of tax reductions are the Revenue Act (1964), Economic Recovery Act (1981), Tax Reform Act (1986), Economic Growth and Tax Relief Reconciliation Act (2001), Jobs and Growth Tax Relief Reconciliation Act (2003), and the American Recovery and Reinvestment Act (2009). ↩
Statistics cited refer to the period 1988-2013 for which consistent data is reported by the Bureau of Labor Statistics. ↩
The measure of total output is drawn from the U.S. Bureau of Labor Statistics prototype total economy production accounts. Unlike the more commonly cited GDP, this measure includes an estimate of the output of capital owned by governments and nonprofit institutions. Source: Harper, Michael, Brent Moulton, Steven Rosenthal, and Dave Wasshausen, “Integrated GDP‐Productivity Accounts,” American Economic Review 99, no. 2, (2009): 74‐79, available at: http://www.aeaweb.org/articles.php?doi=10.1257/aer.99.2.74. ↩
The figures do not add up exactly because of rounding. ↩
The capital intensity contributions for 2010-2014 are 0.1%, -0.4%, -0.3%, 0.0% and -0.1%. Source: Bureau of Labor Statistics. ↩
Andrew B. Abel, "The effects of a baby boom on stock prices and capital accumulation in the presence of social security," Econometrica 71, no. 2 ( 2003): 551-578, available at: http://onlinelibrary.wiley.com/doi/10.1111/1468-0262.00417/epdf. ↩
Year,NNSR 1960,17.04159 1961,16.55058 1962,17.92973 1963,18.53265 1964,19.30692 1965,20.40756 1966,20.44288 1967,18.56129 1968,18.29574 1969,17.76074 1970,15.20867 1971,15.99552 1972,17.00469 1973,18.86042 1974,15.88985 1975,13.2737 1976,15.19264 1977,16.05622 1978,17.39253 1979,17.39788 1980,14.12045 1981,14.83573 1982,11.19921 1983,11.68973 1984,14.8411 1985,13.6992 1986,12.69785 1987,12.07322 1988,12.22502 1989,12.31434 1990,11.02927 1991,9.30826 1992,9.74456 1993,10.00647 1994,11.05984 1995,11.00301 1996,11.78542 1997,12.74744 1998,12.88387 1999,12.78097 2000,11.75416 2001,9.83393 2002,8.75748 2003,8.44529 2004,8.89987 2005,8.76724 2006,8.04013 2007,7.59111 2008,4.83984 2009,2.40354 2010,3.85464 2011,4.18914 2012,5.60557 2013,6.3926 2014,6.8432
Year,P/NNP,P/NNP Trend,G/NNP,G/NNP Trend 1960,65.19,61.18,17.76384535,19.10978292 1961,65.33,61.47,18.11769448,19.08409956 1962,63.7,61.76,18.37039806,19.0584162 1963,63.17,62.04,18.29764643,19.03273283 1964,62.76,62.33,17.93522933,19.00704947 1965,61.89,62.62,17.70045386,18.9813661 1966,61.23,62.91,18.32229581,18.95568274 1967,61.74,63.19,19.69617601,18.92999938 1968,61.65,63.48,20.05507663,18.90431601 1969,62.16,63.77,20.07546332,18.87863265 1970,64.23,64.06,20.56054997,18.85294929 1971,63.47,64.34,20.53580127,18.82726592 1972,62.86,64.63,20.13981064,18.80158256 1973,62.08,64.92,19.05629008,18.7758992 1974,64.45,65.2,19.65837137,18.75021583 1975,66.34,65.49,20.38377492,18.72453247 1976,65.47,65.78,19.33708415,18.69884911 1977,65.1,66.07,18.84592112,18.67316574 1978,64.5,66.35,18.10920072,18.64748238 1979,64.92,66.64,17.6806664,18.62179901 1980,67.47,66.93,18.41178613,18.59611565 1981,66.77,67.22,18.39205452,18.57043229 1982,69.37,67.5,19.42610284,18.54474892 1983,69.22,67.79,19.09012104,18.51906556 1984,67.01,68.08,18.14736117,18.4933822 1985,67.87,68.37,18.43443808,18.46769883 1986,68.54,68.65,18.76226146,18.44201547 1987,69.27,68.94,18.65714698,18.41633211 1988,69.53,69.23,18.24656681,18.39064874 1989,69.5,69.52,18.18875867,18.36496538 1990,70.48,69.8,18.48854217,18.33928202 1991,71.65,70.09,19.03687553,18.31359865 1992,71.55,70.38,18.70176627,18.28791529 1993,71.81,70.66,18.18489614,18.26223192 1994,71.26,70.95,17.67723435,18.23654856 1995,71.58,71.24,17.42009132,18.2108652 1996,71.3,71.53,16.91345601,18.18518183 1997,70.69,71.81,16.56498979,18.15949847 1998,70.8,72.1,16.31504021,18.13381511 1999,70.84,72.39,16.3824832,18.10813174 2000,71.85,72.68,16.39738859,18.08244838 2001,73.13,72.96,17.03603246,18.05676502 2002,73.61,73.25,17.6356258,18.03108165 2003,73.73,73.54,17.82226463,18.00539829 2004,73.36,73.83,17.7442938,17.97971493 2005,73.56,74.11,17.67178672,17.95403156 2006,74.23,74.4,17.72895016,17.9283482 2007,74.5,74.69,17.9084035,17.90266484 2008,76.25,74.98,18.90765694,17.87698147 2009,77.58,75.26,20.01491632,17.85129811 2010,76.42,75.55,19.7219442,17.82561474 2011,76.8,75.84,19.00945628,17.79993138 2012,76.04,76.12,18.35827573,17.77424802 2013,75.95,76.41,17.66019873,17.74856465 2014,75.96,76.7,17.19734939,17.72288129
Year,NNSR,NDIR 1960,0.170415883,0.155276907 1961,0.165505844,0.154574768 1962,0.179297328,0.16408148 1963,0.185326491,0.165457441 1964,0.193069209,0.16636528 1965,0.204075643,0.175340393 1966,0.204428808,0.180187638 1967,0.185612886,0.166448402 1968,0.182957375,0.162356322 1969,0.177607369,0.159471757 1970,0.152086727,0.139079852 1971,0.159955188,0.143886995 1972,0.170046899,0.151579506 1973,0.188604228,0.159211464 1974,0.158898542,0.145413047 1975,0.132736957,0.111240098 1976,0.15192637,0.132338552 1977,0.160562187,0.14780791 1978,0.173925284,0.161609274 1979,0.173978804,0.160975184 1980,0.141204517,0.134010604 1981,0.148357257,0.143794403 1982,0.111992133,0.111548188 1983,0.116897293,0.118759431 1984,0.148410957,0.157870556 1985,0.136992022,0.148069818 1986,0.126978522,0.141557747 1987,0.120732247,0.139549978 1988,0.122250217,0.130672839 1989,0.123143364,0.12793001 1990,0.110292735,0.11600195 1991,0.093082567,0.097829178 1992,0.097445595,0.100682624 1993,0.100064668,0.104738366 1994,0.110598375,0.115724072 1995,0.110030137,0.113028919 1996,0.117854227,0.118027602 1997,0.1274744,0.126476118 1998,0.128838671,0.131048723 1999,0.127809676,0.134766921 2000,0.117541641,0.134816917 2001,0.098339287,0.112970804 2002,0.087574754,0.106856044 2003,0.084452927,0.108704038 2004,0.088998709,0.119232084 2005,0.087672411,0.12544291 2006,0.080401273,0.124538706 2007,0.075911143,0.111112012 2008,0.048398365,0.089605019 2009,0.024035439,0.04912551 2010,0.038546384,0.064830164 2011,0.041891407,0.068101758 2012,0.056055734,0.078514424 2013,0.063925998,0.079176791 2014,0.068432036,0.082478388
Year,NFIR 1960,-1.5 1961,-1.1 1962,-1.5 1963,-2.0 1964,-2.7 1965,-2.9 1966,-2.4 1967,-1.9 1968,-2.1 1969,-1.8 1970,-1.3 1971,-1.6 1972,-1.8 1973,-2.9 1974,-1.3 1975,-2.1 1976,-2.0 1977,-1.3 1978,-1.2 1979,-1.3 1980,-0.7 1981,-0.5 1982,0.0 1983,0.2 1984,0.9 1985,1.1 1986,1.5 1987,1.9 1988,0.8 1989,0.5 1990,0.6 1991,0.5 1992,0.3 1993,0.5 1994,0.5 1995,0.3 1996,0.0 1997,-0.1 1998,0.2 1999,0.7 2000,1.7 2001,1.5 2002,1.9 2003,2.4 2004,3.0 2005,3.8 2006,4.4 2007,3.5 2008,4.1 2009,2.5 2010,2.6 2011,2.6 2012,2.2 2013,1.5 2014,1.4
Year,Capital-labor ratio,Output-labor ratio 1987,0.59,0.68 1988,0.60,0.69 1989,0.60,0.71 1990,0.62,0.71 1991,0.64,0.71 1992,0.66,0.72 1993,0.66,0.73 1994,0.66,0.73 1995,0.66,0.74 1996,0.68,0.76 1997,0.69,0.76 1998,0.70,0.78 1999,0.72,0.80 2000,0.75,0.81 2001,0.79,0.84 2002,0.82,0.87 2003,0.85,0.91 2004,0.86,0.94 2005,0.88,0.98 2006,0.89,0.98 2007,0.91,0.99 2008,0.94,0.99 2009,1.00,1.00 2010,1.01,1.02 2011,1.00,1.07 2012,1.00,1.06 2013,1.00,1.07
Year,Labor efficiency,Technology,Capital intensity,Sum of components (labor productivity) 1988,0.0037,0.0104,0.0022,0.0163 1989,0.0034,0.0248,0.0027,0.0309 1990,0.0032,-0.0096,0.0107,0.0043 1991,0.0086,-0.0295,0.015,-0.0059 1992,0.0076,0.0075,0.008,0.0231 1993,0.0024,0,0.001,0.0035 1994,0.0031,0.0087,-0.0007,0.0111 1995,-0.0005,0.0094,0.0028,0.0117 1996,0.0021,0.0096,0.0083,0.02 1997,0.0012,0.0038,0.0046,0.0097 1998,0.0038,0.0079,0.0087,0.0204 1999,0.0033,0.0077,0.0104,0.0214 2000,0.0001,0.0098,0.012,0.0219 2001,0.0027,0.0104,0.0175,0.0306 2002,0.0043,0.0142,0.0149,0.0334 2003,0.0026,0.0306,0.0118,0.045 2004,0.0016,0.0239,0.0062,0.0317 2005,0.0001,0.0329,0.0065,0.0394 2006,0.0002,0.0026,0.0058,0.0087 2007,0.0035,-0.0087,0.0087,0.0035 2008,0.0034,-0.0198,0.0147,-0.0017 2009,0.0055,-0.0208,0.03,0.0147 2010,0.0021,0.0171,0.0033,0.0225 2011,0.0032,0.045,-0.0018,0.0465 2012,0.0024,-0.015,-0.0023,-0.0149 2013,0.001,0.0137,-0.0001,0.0146