Articles

Working paper 48 - Is Older Better? Maize Hybrid Change on Household Farms in Kenya

 

Working paper 48 - Is Older Better? Maize Hybrid Change on Household Farms in Kenya

Author(s):  Melinda Smale and John Olwande


Introduction:

Kenya has been touted as global maize “success story” for decades (Gerhart 1975; Byerlee and Eicher 1997; Smale and Jayne 2010). Released on the eve of independence, H611, Kenya’s first maize hybrid, a unique, varietal hybrid with Ecuadorean and Kenyan parentage, diffused “at rates as fast or faster than among farmers in the U.S. Corn Belt during the 1930s-1940s”(Gerhart 1975: 51). Paradoxically, policy researchers have more recently lamented that earlier gains in maize productivity have not lived up to their potential (Karanja 1996; Lynam and Hassan 1998; De Groote et al. 2005). Rates of growth in maize production have not kept pace with demand, in large part driven by population growth, so that the country’s import bill has risen during recent years (Kirimi et al. 2011).

The perception of stagnating maize productivity is generally supported with reference to FAO data, although data based and repeated surveys of a panel of farmers (Tegemeo, from 1997) do indicate yield increases. Disagreement among data sources could reflect different spatial representation, especially as maize growing expanded into more marginal areas for production, or differences in temporal representation, since weather conditions are variable under Kenya’s rain-fed production conditions.

Numerous explanations have been advanced for stymied progress. For examples, breeders may have failed to surpass the quality of earlier releases, thwarting gains in yield potential of maize hybrids (Karanja 1996); rising population densities in rural areas may have created inefficient farm size, exacerbating a long-term, secular decline in soil fertility (Lynam and Hassan 1998; Byerlee and Heisey 1997); economic liberalization probably generated uncertainty; and seed liberalization has been partial, curtailing the availability of improved hybrid seed (De Groote et al. 2005). Years ago, Gilbert et al. (1993) pointed out that reported yields understate progress made in counteracting yield losses due to biotic and abiotic stresses through maize improvement (gains from maintaining yields, as compared to augmenting yield potential). Ariga and Jayne (2010) point out that changes in the proportion of intercropped land cause FAO data on maize yields, which is drawn from official MOA estimates, to be biased downward.

This paper begins an effort to disentangle the causes and consequences of Kenya’s maize productivity dilemma by focusing on one component: the age of hybrids grown on farms. Most improved maize seed grown in Kenya has been hybrid. By a hybrid’s “age” we mean the number of years the hybrid has been grown by farmers since its initial year of introduction. Kenyan farmers generally have a long experience with hybrid seed, although they may not choose to grow a hybrid each year. For example, Tegemeo 2010 survey data confirms that on average, farmers began growing improved maize in 1991, with a modal year of 1980. The earliest year among respondents was 1958, and only 4 percent had never grown improved maize. Recently, in a comprehensive analysis of Tegemeo’s panel data, Suri (2011) concluded that farmer learning processes had little to do with whether a farmer chooses to grow a hybrid in any particular year, given the long experience of farmers with hybrid seed in the major maize-growing zones of Kenya.

We argue that it is not adoption of maize hybrids per se which determines the effect of hybrid seed on maize productivity in Kenya today, but replacement of old by new hybrids. Obsolescence of germ plasm is one reason why replacing one hybrid or modern variety by another, and not just replacing its seed, is thought to be necessary for yield progress. For example, this “second stage” of adoption contributed a large proportion of the total economic gains from use of modern wheat seed during and after the Green Revolution in Asia (Byerlee and Traxler, 1995). Slow change of wheat varieties grown by farmers has offset the positive productivity effects of diversifying the genetic base of wheat breeding during the post-Green Revolution period in Punjab, India (Smale et al. 2008).

Based on a 1992 national survey, Hassan (1998) found that the area-weighted average age of all modern varieties grown by farmers (improved open-pollinated and hybrids) was 23 years, although it was only 10 years among hybrid growers, who were concentrated in the higher potential areas. To compare Kenya once again with the US, recent analyses by Magnier et al (2010) indicated that the average “survival” of a maize hybrid on the seed market was only 5 years, and the market share of the typical hybrid peaks at 2 to 3 years.

In this paper, we explore the age of maize hybrids on farms in Kenya and its determinants. In the next section we summarize contextual data on maize yields, use of maize hybrids, and ratios of input to output prices, which are a major determinant of the on-farm profitability of using hybrid seed (Heisey et al. 1998). We then present the data and analytical methods we use to describe and explain hybrid age on farms. Results are presented in the fourth section, followed by a concluding section and policy recommendations.

Is Older Better? Maize Hybrid Change on Household Farms in Kenya