subscribe : Posts | Comments

How much capital is needed to clear the runaway to profitability?


Financial decisions are no different than deciding whether to sign a particular ballplayer or whether a drug produces benefits. We need objective metrics that deal with uncertainty. Uncertainty error will always remain but as put by statistician G.A. Barnard “while there must always be uncertainty involved in statistics, this need not imply any lack of precision the uncertainty may be capable of precise quantitative assessment”. Such metrics are the only alternative to intuitive decisions. And a wide body of research demonstrates that when it comes to uncertainty, human intuitive understanding is often worse than guessing. There are apparently circumstances where experts making decisions based on vast experience yet no systematic analysis can perform worse than flipping a coin.

We are talking about the value of applying statistical analysis to the financial arena in specific in startup or early stage Biotech companies with simply a marketing plan and some clinical research albeit good research, the question becomes “how much capital do we need to raise to bring the technology to profitability”, without having a solid indicator on how long before technological acceptance will gain traction in the market. Only the scientific method produces information about empirical reality using behavior modeling for a technology as it may be absorbed into the current production platform replacing older technologies. Statistical metrics are the only acceptable language of proof.

Most consultants for quantitative Econometrics appreciate the irreplaceable value of statistics. In a 2005 article by John Ioannidis wrote “Why Most Published Research Findings Are False”, a most highly referenced papers, it brought to popular attention to a simmering statistical problem. A December 2, 2011 Wall Street Journal article described how a few years earlier researchers published a study claiming they had destroyed cancer tumors by targeting a specific protein. This got attention, including that of the biotechnology firm Amgen Inc. Amgen, eager to develop a drug based on these findings assigned dozens of researchers try to replicate the findings. But the findings did not replicate after six months of expensive work. This occurrence was just the tip of the iceberg. Most findings in the most prestigious peer-reviewed journals do not replicate. Amgen wound up demonstrating that of 53 “landmark” oncology studies only 6 could be replicated (a batting average of .113!).

But there are two sources of good news concerning the very nature of statistics that will lead the way out of this morass. One is non-reproducible results (i.e. wrong conclusions) using statistics do not occur in a vacuum. One is “if it seems too good to be true”. But the king of red flags is the ubiquitous “small but statistically significant effect”, which is an unintentional way of saying “I guarantee that my finding will not replicate”. So while we may be faced for the time being that only 25% of academic pre-clinical research will replicate, as team at Bayer Health Care in Germany found, we are by no means left to simply guessing which 75% will not replicate. The same things holds true as we progress to phase II and phase III trials. Statistical findings that will not replicate leave clues.

The second item of good news is that when statistics is applied to financial decisions, we have a built in mitigation against the effects of the problems outlined above. Statistical analysis is a perfect fit for financial decisions because in financial decisions an error rate, even an error rate that would be disastrous in other areas, is more than acceptable. Private equity firms raising capital can add a tremendous amount of positive information capital before going to the market by getting a better picture of the length of the runway and how much money will be needed to reach profitability.

Written by Alexander Nussbaum PhD and edited by ken Peters PhD

Leave a Reply

Your email address will not be published. Required fields are marked *