Longitudinal Multilevel Modelling of unbalanced data in R

I'm doing a longitudinal multilevel model analysis of change to look at the impact racism has on health.

My level 2 is thus individual, while my level 1 is time/wave.

I'm using 5 different waves of the UKHLS and in order to include more people I decided to use an unbalanced dataset as every book I read said that MLM can handle that without a problem. However now while trying to fit even one of the simplest models I'm already encountering a problem getting an error message. This is my code and the error message I receive:

m1 <- lmer(data = lusl, generalhealth ~ 1 + wave0 +(1 + wave0 | pidp), na.action=na.exclude)

Error: number of observations (=30962) <= number of random effects (=31520) for term (1 + wave0 | pidp); the random-effects parameters and the residual variance (or scale parameter) are probably unidentifiable
Is there anything I can do about this?