Talent Allocation to Startups versus Incumbents: A Cross-Country Experiment for Germany and the United States

This study implements an online job-choice experiment to measure workers’ preferences over employment in different types of firms. Participants repeatedly choose between hypothetical job offers that vary in wages and selected non-wage working conditions. A key experimental feature is the randomized framing of firm type, which allows us to estimate whether workers assign an independent value to working in startups compared to established firms (or closely related firm-type dimensions). Based on participants’ choices, we estimate WTP for firm type and for non-wage attributes. We further elicit respondents’ expectations about wages and working conditions across firm types. Together, these measures allow us to quantify how preferences and beliefs shape the attractiveness of startups relative to established firms, and how this differs between Germany and the United States.

This pre-analysis plan refers to an online job-choice experiment designed to elicit workers’ preferences for employment in startups versus established firms, as well as their willingness to pay (WTP) for specific non-wage job attributes. The experiment is conducted in Germany and the United States among employed, highly educated workers.

The experimental design consists of a discrete-choice module in which respondents repeatedly choose between two hypothetical job offers that differ in wages and selected job characteristics. We focus on a set of non-wage amenities that are particularly relevant for work in young and innovative firms, including job stability, overtime requirements, opportunities for skill acquisition, team size, work organization, remote work options, multitasking intensity, and perceived societal impact of work. In addition to these amenities, we experimentally vary the firm type. Respondents are randomly assigned to one of two framing conditions: (i) comparing Startups vs. Established Firms, or (ii) comparing firms along the dimensions of Firm Age (Young vs. Old) and Innovativeness (Innovative vs. Not Innovative). This allows us to identify whether workers assign an intrinsic value to the “startup” label beyond observable attributes such as innovativeness or firm maturity.

The choice data enable us to estimate workers’ willingness to pay for each job attribute, including the valuation of startup employment relative to established firms. These preference-based measures are complemented by two additional elicitation modules: (i) expected wage offers at startups and incumbents, both at hiring and after three years, and (ii) perceptions of typical working conditions in each type of firm. Together, these modules allow us to separately assess (a) the utility workers derive from job attributes, (b) their beliefs about wage and amenity differentials between startups and incumbents, and (c) the implied compensating wage differential required to attract skilled workers to startups.

The resulting data allow us to quantify cross-country differences in the attractiveness of startups and to assess whether differences in labor supply to startups are driven by preferences, perceived amenities, or wage expectations.

2025-11-13

2025-12-12

1. WTP to work at a startup (vs. an established firm), estimated from a discrete choice experiment (10 binary job choices per respondent).
2. Expected differences in wages and working conditions between startups and established firms.
3. Implied compensating wage differential for workers to be indifferent between working in a startup and an established firm, computed by combining primary outcomes 1 and 2.

Each choice is between two hypothetical jobs that differ in wage and two amenities. In total, we consider firm type plus eight additional amenities (job stability, necessity of overtime, remote work option, work organization, team size, skill acquisition, meaning of work, and multitasking requirements). Perceived amenity differences by firm type are elicited for a random subset of amenities per respondent and aggregated across the sample.

Heterogeneity in WTP and in implied compensating differentials by: country (Germany vs. U.S.), gender, age, STEM field, migration background, risk tolerance, patience, current job-loss concerns, own startup experience, strength of entrepreneurial orientation in the respondent’s environment, and East vs. West Germany (if sample size permits).

We conduct a stated-preference job-choice experiment that allows us to estimate workers’ WTP for different job characteristics and for employment in different types of firms. Each respondent first answers a short module about their current employment situation. We use this information to anchor hypothetical job offers relative to the respondent’s current job.

Participants are then presented with a series of ten binary job-choice tasks. In each task, respondents choose between two hypothetical jobs that differ in monetary compensation and a small subset of non-wage job characteristics. To reduce cognitive load, only a limited number of characteristics vary within each choice pair to differences in wage and two amenities. Wage levels are directly tied to the respondent’s reported current wage, with proportional upward or downward variation within a bounded range.

A key feature of the experiment is the randomized framing of firm type. Respondents are randomly assigned to one of two framing conditions that describe either (i) a “startup vs. established firm” comparison or (ii) a 2×2 framing with Young/Old and Innovative/Not Innovative. This randomization enables us to identify whether respondents place intrinsic value on working in startups beyond the specific job attributes themselves.

After the discrete-choice module, we elicit expected wages by firm type (at hire and after 3 years) and perceived amenity differences by firm type, and collect demographics/background information. These additional measures allow us to compare preferences to beliefs, evaluate heterogeneity across worker groups, and assess how preferences and perceptions jointly shape the attractiveness of startups relative to established firms.

Randomization by a computer

1. Between-subjects: firm-type framing (Group A vs. Group B).
2. Within-subject: randomized attribute profiles across the 10 binary choice tasks.

No

1. 10,000 individual choices, 5,000 individuals per country
2. 100,000 individual choices over hypothetical jobs

10,000 individuals × 10 choices = 100,000 choice observations.

1/3 of the sample receives the startup vs. established firm framing, 2/3 receive the framing with firm age and innovativeness.
≈3,333 individuals (≈33,330 choices firm-type framing Group A).
≈6,666 individuals (≈66,660 choices firm-type framing Group B).