Actually, Operation Iraqi Freedom has shown that anti-jamming capability would be important for all PGMs in the future. USAF made it sounds so easy to counter the Russian-made GPS jammers in Iraq, and revealed the ironic fact that one of the six Iraqi jammer was destroyed by GPS weapon. But guess what? In 2003, USAF asked Boeing to develop anti-jamming devices for weapons such as JDAM, and asked how fast it could be done. Ultimately they did not have to rush anything new in service during wartime, but jamming capability is one of the technical "lesson-learned" that came out of OIF.
Improving accuracy is "relatively" easy, even if it may not be the most cost effective in the short run. You could use more expensive guidance devices to achieve the accuracy. By comparison, ECM and anti-jamming capability is difficult. If the technology is not there yet, spending more money now would not give you the capability now.
Taiwan's lack of strategic depth is an extremely serious operational drawback to say the very least. But here, this also indicates smaller cover area, for missile defense or jamming defense measures. A smaller number of strong jammers combine with large quantity of expendable jammers could be cost-effective.
Of course, nothing says it would be easy for ROCAF. One obvious problem is that PLA has multiple guidance sources, GPS / GLONASS now and in the future indigenous and EU satellites.
Production cost is certainly important, but so is ECM. I would not dismiss the jamming issue so easily against the cost equation. In fact anti-jamming capability will also affect the overall cost and cost effectiveness.