How do we explain the rejection of argument 1?

[harishk05]

We may have already discussed this issue, so apologies if this is a duplicate issue/post.

Thinking back on the Informs presentation, there were two arguments. Argument 1 was that QuickPay should reduce project delays. Argument 2 was that QuickPay should increase project delays for financially constrained firms.

The empirical results show that argument 1 does not hold. In fact, QuickPay increases project delays (for all firms). The empirical results also show support for argument 2: QuickPay increases project delays for financially constrained firms.

But back to argument 1: not only does argument 1 not hold, the empirical results show that there is something going on that leads to the opposite effect of that predicted by argument 1. All firms (not just financially constrained ones) face a delay. By argument 2, the financially constrained firms face an additional delay. But how do we explain empirical result 1? Do we need to expand on this? Will one of our other theories help explain what is going on?

Comments?

[JNing0]

Hi Harish, we do have empirical evidence that supports Theory 1, from the 2014-implementation of QuickPay on large businesses. In fact, that is our only empirical support for Theory 1.

Your comment is an important reason that argues for including the 2014 QuickPay in the analysis. I discussed the need for including large businesses in this file. Leaving aside the specifics such as which model to use, I think the 2014-QuickPay should be in the paper.

[vob2]

QuickPay increases delays to all small firms. We can argue that all small firms are financially constrained, just to a different degree. The variation in that degree is captured by CF variables.

So, not surprisingly, all small firms experience delays. That does not mean our first force is not at work. It is just all second force is more powerful and we do not have a proxy that captures variation in the first force.

I agree with Jie, the second implementation of QuickPay is the one that can and does give us a glimpse at financially unconstrained firms, and turns off the second force. That does not necessarily mean that a priori we will see shorter project durations because there are other forces that could delay projects after QuickPay. The fact that we see shorter project durations for large firms supports the first force, though.

If we are careful in explaining our empirical setup for the second QuickPay implementation (so that we do not ask readers to perform mathematical operations mentally when looking at our tables), we should include it. Because setup will likely be different from the first implementation, we probably will have this as a separate table (or separate panel in the same table) and separate discussion.

[harishk05]

Thank you both. This is helpful. I need to give it some more thought, but I see the outline of the argument. I also agree that this provides a rationale for including the second implementation of QuickPay (and I also agree that we will have to take care to present it carefully).