Does GDP distort Mexico's economic performance?

• Author: Fuess, Scott M., Jr.
• Position: Gross domestic product

1. Introduction

   Between 1961 and 1980, Mexico's real gross domestic product (GDP) per capita grew nearly 3.5% annually. In early 1982 the Mexican government announced that it could not repay its debt obligations. Following this debt crisis, the country's GDP growth slumped. In fact, over the course of the 1980s, GDP per capita shrank nearly 0.5% per year. After this collapse the Mexican government revamped its policies, focusing on deregulation, privatization of many state enterprises, and trade liberalization. As the policy makers shifted course, economists have tried to identify factors that caused the economy's dramatic slump.(1)

   In focusing on what may have caused Mexico's decline, few economists or policy makers seem to have questioned the accuracy of the GDP figures they sought to explain. This oversight is curious, considering the well-known view that national income and product accounts are likely to present a distorted impression of a developing economy.(2) Perhaps the decline was so sudden and obvious that few observers felt compelled to assess how accurately GDP portrayed the economy's actual performance before and after the debt crisis.

   GDP does not measure output reliably because it includes not only the final output produced by an economy's markets, but also transactional activities, which are intermediate to production (see Wallis and North 1986; North 1987). Transactional activities involve outlays to support exchange, not actual output consumed. As an economy develops, specialization causes market transactions to multiply. This increased interdependence begets economic growth by spurring innovation, better resource allocation, and economies of scale. Greater interdependence produces a wider variety of output. Nevertheless, North (p. 419) has reminded economists that exchange is costly: The so-called Walrasian auctioneer must be paid.

   It is costly for government to provide national defense, administer justice, and enforce regulations. Financial intermediaries charge fees to reallocate purchasing power and arrange insurance coverage. Wholesalers and retailers are paid to bring buyers and sellers together. These transactional activities are intermediate inputs, yet they are included in standard measures of domestic product. Therefore, GDP exaggerates output because it includes the actual goods consumed and outlays to support exchange of those goods. This inclusion causes GDP to overstate the level of output, but the impact on output growth is unclear. If transactional activities occupy an expanding (shrinking) share of GDP over time, then measured GDP will overstate (understate) economic growth.

   Transactional activities should have varied in Mexico over the last three decades. Structural transformation of the economy - expansion of manufacturing and services, the decline of traditional agriculture - increased economic interdependence. Government policy shifts may have affected transaction costs. The Mexican government pursued a policy of stabilizing development in the 1960s, only modestly increasing outlays and regulation. In the 1970s government expanded regulatory activities, spending, and foreign borrowing.(3) With the reforms enacted after the 1982 debt crisis, government's direct participation in the economy diminished. Yet the turbulence of the 1980s, which featured accelerating inflation and exchange rate instability, should have led Mexicans to search more actively for assets that retained real value, raising the costs of arranging financial transactions.

   Income and product accounts can misstate an economy's performance not only because they include transactional activities as part of output, but also because they fail to register some types of production. GDP reflects neither home production (housework) nor goods and services generated outside households in the informal sector. There is reason to believe that these types of unrecorded production are a substantial part of Mexico's society. The Centro de Estudios Economicos del Sector Privado (1986) estimated that in the 1980s, the informal sector had grown to as much as 25% of the country's measured GDP. An OECD survey (1992, p. 55) of Mexico reported that informal activities accounted for about one-third of all employment during the 1980s. The survey reported (p. 219) that "as economic activity contracted in the 1980's . . . the share of informal activities widened in total employment." If household/informal production grows relative to GDE then measured GDP will understate the economy's actual growth.

   With variations in transactional activities and unrecorded production likely, this study recalculates Mexico's output growth. Fuess and Van den Berg (1996) have shown that transactional activities and nonmarket production combine to overstate expansion of the United States economy. Using their calculation methods as a guide, we adjust Mexico's GDP to generate actual output growth per capita, which we compare with that of per capita GDP.

   To account for transactional activities, we needed disaggregated GDP data, which are available since the 1960s. To measure Mexico's stock of labor, which is necessary to calculate actual output, we used decennial census data collected in 1960, 1970, 1980, and 1990. Given the availability of data, our analysis covers the 1961-1990 period.

   [TABULAR DATA FOR TABLE 1 OMITTED]

   We find that GDP does not accurately reflect Mexico's actual output. In the 1960s, the economy expanded more quickly than GDP implies. In the 1970s, in contrast, growth was much worse than GDP suggests. The economy slumped in the 1980s, but not as terribly as the official figures indicate. Mexico's economy did not collapse suddenly in the early 1980s: Actual economic growth had already slowed dramatically in the 1970s.

2. Growth of GDP Per Capita

   Table 1 reports the growth of Mexico's real GDP per capita (see Appendix: GDP Per Capita). Over the full 1961-1990 sample period, it increased at a mean rate of 2.16% per year. This 30-year stretch masks distinct variations in the economy's performance.

   In the 1960s there was steady development with low inflation. Real GDP per capita grew at an average annual rate of 3.57% (see Table 1). The 1970s began and ended with GDP growth but there was a slump mid-decade, when inflation started to accelerate. Yet growth still averaged 3.39% per year which is nearly the same rate as for the 1960s. GDP was still expanding as the 1980s began but entered a prolonged contraction after the 1982 financial crisis, with growth resuming only at the end of the 1980s. Over this stormy decade, Mexico's GDP per capita shrank 0.49% per year. Given these distinct phases, we also analyze the three decade-long subperiods 1961-1970, 1971-1980, and 1981-1990.(4)

   According to the official GDP figures, Mexico's economy expanded during the "populist" 1970s almost as rapidly as in the 1960s. Following the debt crisis of the early 1980s, output per head shrank, despite economic reforms. To see whether this is an accurate impression of the economy's actual performance, we recalculated per capita output growth. We controlled first for transactional activities and then for nonmarket production.

3. Adjusting for Transactional Activities

   To calculate the final output produced by an economy's market sector, it is necessary to adjust GDP for transactional activities, which are inputs in the production process. Transactional activities are those operations that secure property rights and support market-based exchange. In their analyses of the United States, Wallis and North (1986) and Fuess and Van den Berg (1996) distinguished between public and private sector transactional activities. In the public sector, they focused on defense, outlays for protection of persons and property, and expenditures for government administration. In the private sector, they concentrated on financial and marketing activities.

   With aggregate time series data, there is seldom a clear distinction between transactional activities, the market process, and the final output exchanged.(5) Using the studies by Wallis and North and Fuess and Van den Berg for guidance, we concentrated on broad categories of activity that include transactional functions. Then we excluded them from reported GDP to derive Mexico's transactions-adjusted GDP.

   Controlling for Public Sector Transactional Activities

   Many governmental functions establish or protect property rights. These efforts include national defense, the legal system, and public safety. Other endeavors, such as regulation, general administration, and legislative activities, use real resources to maintain social institutions or alter market outcomes. These efforts may help support the production of goods, but Eisner (1989, p. 40) has argued that a measure of final output should exclude these intermediate activities. To account for government transaction costs (GTRANS), we subtracted them from real GDP (see Appendix: Public Sector Transactional Activities).

   GTRANS varied little during the 1960s - between 2.57 % of GDP (1961) and 2.65 % (1970). The GTRANS share of GDP expanded somewhat in the 1970s and early 1980s, reaching 3.50% (1984) before falling to 2.99% (1990). Because the GTRANS share of GDP varied so little, government transactional activities hardly distort the impression of Mexico's economic growth. Controlling for GTRANS over the entire sample period, average annual improvement in output per capita is 2.14%, virtually identical to the 2.16% rate using GDP (see Table 1). Even for the populist 1970s, a GTRANS control cuts only 0.04 points off the GDP growth rate.

   Controlling for Public and Private Sector Transactional Activities

   Activities to support exchange also occur in the private sector. Financial intermediaries are paid to reallocate purchasing power, preserve wealth, and spread risk. These transactional functions occur in the finance-insurance (FINS) sector of Mexico's economy. Parties value...