There has been increasing interest recently in hypothesis testing with ineq
uality restrictions. An important example in time series econometrics is hy
potheses on autoregressive conditional heteroskedasticity (ARCH). We propos
e a one-sided test for ARCH effects using a wavelet spectral density estima
tor at frequency zero of a squared regression residual series. The square o
f an ARCH process is positively correlated at all lags, resulting in a spec
tral mode at frequency zero. In particular, it has a spectral peak at frequ
ency zero when ARCH effects are persistent or when ARCH effects are small a
t each individual lag but carry over a long distributional lag. As a joint
time-frequency decomposition method, wavelets can effectively capture spect
ral peaks. We expect that wavelets are more powerful than kernels in small
samples when ARCH effects are persistent or when ARCH effects have a long d
istributional lag. This is confirmed in a simulation study.