Most of the analyses of the costs of evoting systems have looked at the disparity in initial capital costs as compared to other types of voting systems, such as optical scan voting systems. Those initial capital costs are at least a factor of 2 or 3 higher than for precinct-count optical systems, depending on the average number of voting stations per polling place in a given jurisdiction.

Proponents of evoting systems argue that those initial higher capital costs will be recovered over time via reduced costs for printing paper ballots. owever, the payback times (for paperless e-voting systems) are estimated at 15 to 20 years, which may well exceed the operational lifetime of such systems.

In states, such as California, which will require all e-voting systems to provide an voter-verified paper audit trail, the operating costs will be even higher, and the payback times even longer.

In addition, there are significant hidden costs associated with e-voting machines. For example, the process of conducting logic and accuracy tests (typically required before each election) are significantly more labor intensive than corresponding tests on optical ballot scanners. 

The former require test votes to be entered by hand, while the latter can be tested by simply counting a standard test deck of optical scan ballots. In addition, an e-voting solution requires a significantly higher number of total machines, since each polling place will require 4 to 8 machines, as opposed to a single optical scanner per polling place. he more machines you have, the higher the costs for storage, hardware maintenance contracts, software maintenance contracts, and replacement of parts (e.g., batteries). Once these costs are factored in, along with the costs of consumables for the voter-verified paper audit trail, the operating costs may exceed those of a precinct-count optical scan solution, even when factoring in the costs of the pre-printed optical scan ballots.