In yeast to get secreted they must go through the secretory pathway, from the ER to Golgi to vesicles that fuse with the plasma membrane. To get into the ER they need a signal peptide to be recognized by a cytoplasmic protein, that leads the ribosome to the ER surface so that the growing polypeptide gets inside. In the absence of the signal peptide the protein would not go inside the ER. Maybe the cell has the two options, to make the protein with and without signal peptide, so that it can also be cytoplasmic. I bet the protein folding is not the same inside and outside the cell.

Hello, all well known galactosidase producers are intracellular.

Being intracellular or extracellular is a trait that was chosen during early evolution.

Broadly speaking, if the cell has a sort of membrane transporters that transport lactose from surrounding into the cytoplasm, then it's favorable to keep its β-galactosidase within the cytoplasm. If such transporters are lacking, then the cell is compelled to put that enzyme extracellularly in order to be metabolize lactose. Till now, it is still largely unknown how the bacterial exoenzymes are secreted from cytoplasm to the environment.

However, secreting a protein in bacteria seems complicated and energy-consuming.

In cytoplasm, the enzyme encounters very few (if any) inhibitors such as L-ribose and iodine or other compounds/molecules that might affect its activity. For a bacterial cell to transport a protein (or enzyme) beyond cell membrane and wall, it should have a signaling peptide to be recognized by the corresponding secretory system that carry out the transportation OR the protein is synthesized to the final destination - a process that is quit complicated and energy-consuming.

If the enzyme is to be put in the periplasmic space, then it's synthesized by ribosomes that work in very close proximity with a membrane-bound secretory proteins that coordinate peptide secretion during translation process. How exactly the exoenzymes pass the peptidoglycan is still unknown.