I do not think geom_col takes y limits and you have to scale the data manually as suggested in this stack overflow thread. Here is an example based on that.
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
library(ggplot2)
dat <- data.frame(names = c("A", "B", "C"), Value = c(80, 70, 90))
ggplot(dat, aes(x = names, y = Value)) + geom_col()

OFFSET = 60
dat <- dat %>% mutate(ValueOffset = Value - OFFSET)
ggplot(dat, aes(x = names, y = ValueOffset)) + geom_col() +
scale_y_continuous(labels = function(x) x + OFFSET) +
labs(x = "Model Type", y = "Squared Error Number")

Created on 2019-05-01 by the reprex package (v0.2.1)