经济与社会科学2019papers.ssrn.com
独自幸存：单人创业公司与创始团队（2018）
Sole Survivors: Solo Ventures Versus Founding Teams (2018)
进阶 · 约 50 分钟读完前置知识：创业与融资基础、计量回归、生存分析
导读
论文《Sole Survivors》研究单人创始的创业公司与多人创始团队在表现和存活上的差异，结合公开数据和 Kickstarter 项目做实证比较。其核心结论似乎指向单创始人并不天然更差，但评论者认为数据截尾、注册时点和融资筛选会严重干扰因果解读。
HN 关注点：单创始人是否真的更优，以及这类创业研究中的样本选择、幸存者偏差和因果推断问题。
HN 评论精选
这几条评论里：质疑 2 补充 1 亲历 1
亲历亲历说明单创始人风险集中：创始人个人健康危机可直接导致公司关闭。
作为单一创始人是一项高风险创业，和所有高风险事物一样，它也有潜在收益。我作为单一创始人成功经营公司 5 年，直到有一天被告知我将四肢瘫痪，我不得不关闭公司，因为没有其他高管能接手。我的产品曾进入 Facebook Fb Start 加速器，有 25 万用户；任何表示有兴趣收购我产品的公司都希望我去为他们工作，而我当时正在术后恢复，无法做到。单一创始人有其优点，但当你的健康或个人生活出问题时，你会失去公司。
Abishek_Muthian查看英文原文
Being single founder is of high risk venture, as anything with high risk it has potential gains as well. I ran my company successfully as a single founder for 5 years, till one day I was told I would become a quadriplegic & I had to close my company as there's no other executive to take care of it. I had Facebook Fb Start accelerated product with 2,50,000 customers , any company expressing interest in acquiring my product wanted me to work for them; which wasn't possible as I was recovering from surgery. Being single founder has it's merit, but when things go south with your health or anything personal; you'll loose your company.
质疑质疑样本选择：注册和众筹数据把已跨过门槛的单创始人与早期团队混在一起，Kickstarter 不具代表性。
唉。这项研究没用。他们有两类数据来源：公开可得数据（他们自己说存在严重的截尾问题）和 Kickstarter 项目。人们会在认为值得注册公司时才去注册。这几乎是同义反复，但想想看：单创始人什么时候会注册公司？当注册的好处值得花时间时。可能是融资时、有客户需要限制责任时，或盈利后想要税收优惠时。那双创始人呢？很多情况下第一天就注册，或很早就注册，因为需要合法确立股权。所以你选到的是已有正常业务的单创始人，却选到各个阶段的配对创始人。我认为大多数创业项目可能在零收入、几乎零客户或用户时就失败了。Kickstarter 无论他们怎么想都不会有代表性。
ltbarcly3查看英文原文
Sigh. This study is useless. They have two sources of data, publicly available data (which they say has major censoring problems) and kickstarter projects. People will incorporate at the point that they believe it is worthwhile to incorporate. This is basically a tautology, but just think about it. When would a solo founder incorporate? When the benefit of incorporation makes it worth the time to do it. This might be when they are taking investment, when they have customers and need to limit their liability, or when they are making profits and want tax benefits. What about pair founders? In many cases on the first day, but often very early on because it is necessary to legally establish their ownership stake in the company. So you are selecting solo founders who have a functioning business and pair founders at every stage. I think most ventures probably fail with zero revenue and virtually zero customers and/or users. Kickstarter won't be representative of anything, even in their wildest dreams.
质疑指出声明时点差异与融资偏见造成混杂，单创始人样本存在幸存者偏差，因果难证。
这些效应可能来自无关的假象：1）两人或更多人一起做任务时，可能比个人更早宣布自己是一个组织并自称创始人；个人可能更倾向等到有客户、接近盈利等时才做此声明。2）摘要说“投资者很少、甚至几乎从不资助由单一创业者创办的初创公司”。因此观察结果与单创始人克服这一偏见混杂在一起。这句话也指向幸存者偏差：也就是说，单创始人必须在获得融资前表现优于团队，较弱的单创始人更少能存活到被研究观察到。如果文章结论正确，我不会惊讶。但除非说服创业者参加长达数十年的实验、随机分为单创始人和团队（这几乎不可能获得有代表性的创始人样本同意），否则很难令人信服地证明因果。
sagebird查看英文原文
The effects may be due to irrelevant artifacts: 1.) Pairs or more of people working on a task may declare themselves to be an organization and call themselves founders earlier than an individual. An individual may be more likely to wait until they have a customer, are near profitability, etc to make that declaration. 2.) The summary says that "investors rarely, if ever, fund startups founded by a solo entrepreneur". So the observations are confounded with solo founders overcoming this prejudice. This quote also points to a survivor bias in play. IE, solo founders must have higher performance than teams before getting funded, and less of the weak solo founders survive to be observed by the study. I wouldn't be surprised if the article had a correct conclusion. Though it would be very hard to convincingly show causality short of convincing entrepreneurs to take part in a decades long experiment where they are randomly divided into solo founders and teams (which would be nearly impossible to get a representative sample of founders to agree to).
补充从风险偏好解释：单创始人分布更宽，观察到的多是成功正尾。
我认为个人创始人更可能快速行动、承担更多风险。第二位创始人无论如何都会成为缓和因素。这可能意味着个人创始人在成功潜力上分布更宽。我们主要注意到分布的正尾，因为负尾被截断（一旦失败，很难更失败，而成功相比之下几乎无限）。
bhouston查看英文原文
I think that individual founders are more likely to move fast and take more risks. A second founder will be a mitigating factor no matter what. It may mean that individual founders have wider distributions in terms of their potential for success. And we notice mostly the positive tail of the distribution because the negative tail is cut off (once you fail, it is hard to fail more, where as success is nearly unlimited in comparison.)
相关资源
从这篇论文所有 HN 讨论帖的全部评论里整理出来的链接。
历次讨论
- 2019Hacker News 讨论帖187 分 · 33 评
同领域还值得读
一切的回报率，1870–2015
The Rate of Return on Everything, 1870–2015
利用1870–2015年多国的长期数据，估算股票、住房等各类资本的回报率，并将其与经济增长率进行比较。
经济与社会科学2019 · 487 分 · 260 评进阶 · 约 100 分钟编辑精选一切资产的回报率，1870–2015
The Rate of Return on Everything, 1870–2015
构建了 1870–2015 年 16 个国家的长周期资产回报数据集，覆盖股票、住房、政府债券和短期票据，系统比较各类资产的真实回报与波动性。
经济与社会科学2018 · 215 分 · 151 评进阶 · 约 120 分钟编辑精选一切资产的回报率，1870–2015（2019）
The Rate of Return on Everything, 1870–2015 (2019)
这篇发表于《经济学季刊》的论文构建了 1870–2015 年间 16 个主要发达国家的股票、债券、房地产等资产收益率的长周期数据库。
经济与社会科学2024 · 190 分 · 148 评进阶 · 约 80 分钟编辑精选想法是否越来越难找？
Are ideas getting harder to find?
探讨“想法是否越来越难找到”：通过衡量多个领域的“研究生产率”，发现维持相同的增长需要投入指数级增长的研究资源，即研究效率在持续下降。
经济与社会科学2016 · 182 分 · 106 评入门 · 约 60 分钟编辑精选比特币与区块链的经济极限
The Economic Limits of Bitcoin and the Blockchain
芝加哥大学 Eric Budish 的论文从经济学角度分析比特币与区块链的根本限制：核心论点是只要攻击网络（双花或砸盘做空）的收益超过挖矿成本，51% 算力攻击就有利可图，而挖矿…
经济与社会科学2018 · 176 分 · 37 评进阶 · 约 90 分钟编辑精选