A deeper look at asymmetries in UK consumers’ expenditure: the nonparametric analysis of 100 disaggregates

Abstract

The literature on testing for the presence of cyclical asymmetry in UK consumers’ expenditure is extended via the application of nonparametric tests to data subject to a higher degree of disaggregation than considered in previous studies. The results obtained at an intermediate level of disaggregation depict a positive relationship between the durability of goods and the degree of asymmetry they exhibit. At high levels of disaggregation to specific expenditure categories, it is found that the aggregate and intermediate evidence of deepness asymmetry is driven by a relatively small number of expenditures, with almost no evidence of deepness asymmetry in non-durable categories of expenditure. Prominent among the durable and semi-durable good expenditure categories exhibiting significant positive deepness asymmetry are expenditures relating to housing fittings and communication, consistent with a degree of ‘lumpiness’ in expenditure which may be associated with the impact of credit rationing or threshold effects.

Notes

See Psaradakis (2000) for discussion of a number of commonly employed tests of asymmetry and nonlinearity.

In comparison, Speight and McMillan (1997) examine 13 categories of consumers’ expenditure; Cook (2000) examines two categories, while Holly and Stannett (1995) consider the aggregate series alone.

The two aggregate series considered are ‘national’ and ‘domestic’ measures. The series differ as a consequence of the latter's inclusion of net tourist expenditure.

Consumer Trends four character identifiers for these series, in the order listed in the results tables, are as follows: Table 1: ABJR, ZAKW, ZAKY, ZALA, ZAVO, ZAVW, ZAWC, ZAWM, ZAWW, ZAXA, ZAXS, ZAYG, ZWUN, ZWUT; Table 2: UTID, UTIH, UTIL, UTIP; Table 3: ATQX, ATRD, ATRR, ATRV, ATRZ, ATSD, LLKX, LLKY, LLKZ, LLLA, LLLB, LLLC, TMMI, TMML, TMMZ, TMNB, TMNO, UWIC, XYJP, XYJR, XYJT, ZAYM; Table 4: ATQV, ATRF, ATRJ, ATSH, ATSL, ATSX, AWUW, CDZQ, LLLZ, LLMA, LLMB, LLMC, LLMD, XYJN, XYJO, XYJQ, XYJS, XYJU, XYJX, ZAVK; Table 5: ATSP, ATUA, AWUX, CCTK, CCTL, CCTM, CCTN, CCTO, CCTT, CCTU, CCTY, CCUA, CDZY, JRBA, LLLL, LLLM, LLLN, LLLO, LLLP, LLLQ, LTZA, LTZC, TTAB, UTHW, UTXP, UTZN, UUIS, UUVG, UWBK, UWBL, UWFD, UWFX, UWGH, UWGI, UWHO, UWIB, UWKQ, XYJW, ZAKY, ZWUN, ZWUP, ZWUR.

It should be noted that although some components, for example ‘Transport’, appear under durable, semi-durable and non-durable expenditure, in each case expenditure is defined as that part of the component relating to the classification (durable, semi-durable and non-durable) being considered.

All series considered possess clear trends. The results of ADF(5) tests including an intercept and trend show that the unit root hypothesis can be rejected at the 5% level of significance for only three of the 103 series examined. These series are clothing materials; newspapers, books and stationery; electricity, gas and other fuels. In these instances the series are concluded to exhibit trend stationarity, while in all other instances it is inferred that a non-stationary trend is present.

Although numerous detrending procedures exist (see, inter alia, Harvey and Jaeger, 1993; Cogley and Nason, 1995; Christiano and Fitzgerald, 2003) the HP filter has a number of attractive features and advantages over its rivals in the present context. In particular, because of its linear structure, the HP filter will not induce spurious asymmetry in the derived cyclical components. The continued use of the HP filter more generally in the economics literature is noted by Ravn and Uhlig (2002).

Following the suggestion of Christiano and Fitzgerald (2003), two years of observations (8 observations) are omitted at the start and end of the derived trend to overcome the ‘endpoint issue’ (see, inter alia, Kaiser and Maravall, 2001).

Whilst Psaradakis and Sola (2003) show that tests of asymmetry may possess low power when applied to the business cycle component of macroeconomic time, and a higher probability threshold might therefore be employed, they do not examine the Triples test employed here, and the conventional significance level of 5% is retained here. Nevertheless, the potential for low test power should be borne in mind in the results reported here, and the reader may wish to consider significance up to a higher probability threshold of, say, 10% (though this is only pertinent to four instances of test result in Table 5).

It should be noted that this average figure masks substantial variation over the sample. More precisely, while these series account for 7% of the expenditure on semi-durable goods in the initial period, they account for 30.4% of semi-durable expenditure in the final period. This increase is driven primarily by year-on-year increases in expenditure on recreation and culture, and sports equipment. Such variation is not observed for durable and non-durable goods, the reported average figures being representative of all observations.

The significance deepness in the three specific non-durable expenditure categories is less easily reconciled, but it is of note that the instances of significance are approximately the same as the power of the test and the potential for Type I error in significance testing.

See Caballero (1993) for a theoretical analysis of this issue.