Estimating permanent income and wealth of the US farm households

Abstract

Farm households are unique in the way they derive income when compared to nonfarm households. Farm operators and spouses have dual income sources, from farm and off-farm activities. Further, farm households on an average possess higher wealth than nonfarm households. This article estimates permanent income for the US farm households using data from a large national farm level survey. The estimated income is then used to identify the determinants of wealth accumulation by the US farm households. Results confirm that permanent income is closely related to age of the operator, education, occupation, farm size, location and number of earners in the household. Along with age, permanent income is a significant determinant of household wealth. It was also found that the wealth–income curve is nonlinear, upward sloping, and convex. Hausman's specification test indicates that variations in farm household wealth is better explained by estimated permanent income than observed total household income. Off-farm income cannot be treated as residual or transitory income.

Notes

¹ Savings and wealth accumulation are used interchangeably in this article.

² Defined as the anticipated income that could be average of past several years.

³ Defined as the unanticipated income, the difference between the measured income and the permanent income.

⁴ In the estimation process economists treat predicted income as ‘permanent income’ and the residual as ‘transitory income’ see for example Musgrove (1979), Bhalla (1980) and Wolpin (1982).

⁵ See Paxson (1992) for details on these three measures of savings.

⁶ Paxson (1992) used deviations in regional rainfall as a proxy for estimating transitory income. Wolpin (1982) used rainfall distribution to estimate direct and indirect consumption of durables in rural India.

⁷ Kwon et al. (2006) used rainfall variability in counties of Iowa as a proxy for transitory shock to farm income.

⁸ However, a common finding for farm households in the US (Reid, 1952; Friedman, 1957) and in developing countries (Paxson, 1992) is that transitory income shocks do not affect farm household consumption.

⁹ Transitory income or the residual income (Y^T ), the difference between observed total income and predicted income from farming, by definition has a mean of zero over time and therefore Y^T will not have a significant impact on wealth.

¹⁰ In some sense classification of farm under farm typology reflects farm size (going from left to right in Table 1) as measured by gross farm sales.

¹¹ We estimated income equations separately by each farm size but did not find them to be significantly different. Further, pooling test shows that income equations by different categories can be combined. However, to tease out the effects of two income categories we used two dummy variables in the regression analysis. In addition, the National Comission on Small Farms, USDA believes that there are two farm sizes, small (sales <$250 000) and large (sales >250 000).

¹² King and Dicks-Mireaux (1982) outline a measure of permanent income that can be calculated using cross-sectional data.

¹³ To assess the impact on age-earnings profile, age-squared variable was included in the regression but did not find any nonlinear specifications that notably improved the fit over the linear specification.

¹⁴ Last two models are estimated to assess the sensitivity of the estimates.

¹⁵ Residual income is the difference between total household income and predicted income from farming.

¹⁶ Statistically, off-farming (nonfarm) income does not have a mean of zero overtime.