In the picture below, the input of the rectifier is 690VAC max, so the DC output can reach up to 1kVDC. The DC output goes to a low power DC/DC Converter. In the picture on the left, I'm using a single varistor for transient protection. In the picture on the right, I have three TVS diodes in series. Would the two cases accomplish the same task? As far as I know, TVS are faster than MOVs and can absorb less energy.
This would work in an ideal world. However, I'd be concerned that IRL one of those 450V TVSes may have a slightly larger reverse leakage than the others, and thus a much smaller voltage drop. Once one of those starts conducting even a little more than the others, it's a domino effect. You may be able to address it by putting a resistor across each one that will pass far more current than the diode leakage, forming a more reliable voltage divider.
An MOV is not the same as a TransZorb. An MOV has a soft avalanche while TransZorbs have a sharp zener-like cut-off. Both are transient absorbing devices.
For constant voltage clamping zener diodes are better suited. In terms of protecting power supplies it is best to put MOV's or TransZorbs or RF noise filter capacitors (1 nF X2 rated)on the AC side of the diodes. Zener diodes, if used, would be installed after the rectifier diodes.
I worked in the TVS / SPD industry for 15 years, and putting devices in series was common. They are more balanced than you think. As they do not make individual units (so far) with a rating over 1,200 VAC, to protect 4,160 VAC oil-pump motors we put 4 or 5 in series, then several of those in parallel to get the KA rating required by the vendor, and the safety margin required by UL and ISO.