Initial Retention of External Accountants in Startup Ventures

Abstract

This study investigates the determinants of the initial retention of external accountants by new US startups. We test several arguments related to agency and credibility needs, financial activity and complexity, and accounting support and advice. Consistent with financial activity and complexity arguments, we find that intended scale and bank account activity are positively associated with both the actual and intended retention of an accountant. Comparing the decision to retain an accountant with the retention of another professional service, retaining a lawyer, we find many of the same factors to be significant. However, we find that lawyers are more likely to be retained during heightened uncertainty, primarily given the role of lawyers with patent and intellectual property assistance. Overall, these results suggest that some of the determinants of external accountant retention contrast with the determinants of internal accounting resource adoption, which are primarily related to decision-making needs and uncertainty.

Acknowledgements

The authors would like to thank the Ewing Kauffman Foundation for sponsoring the Panel Study of Entrepreneurial Dynamics II. The financial support of the Wharton School and Ernst & Young is greatly appreciated. Data are available from sources identified in the study.

Notes

Examples of ineligibility at this stage included the respondent not living at the household that was randomly selected and incorrect contact phone number provided.

The Pearson (Spearman) correlation between first and fifth year sales is 0.87 (0.86) and the results are not affected by the choice of sales used in the regressions.

To ensure that the intention to retain an accountant results are not driven by the differences in timing between actual and intended action of the four variables listed above, in the regression we separated each variable into whether the startup had: (a) undertaken the activity; or (b) had not undertaken the activity but intended to. In unreported results, the coefficients for each of these four variable pairs were found to be very similar in magnitude and not significantly different.

The opening of a bank account potentially captures more than financial complexity. For example, in the actual retention regression it may also capture venture stage of operation, while in the intention regressions it may capture intended venture scale. To address concerns that ventures that do not intend to open a bank account are inherently different from startups that do, we reran the regressions excluding those startups that have not and do not intend to open a bank account. The results are consistent with the original findings and consequently not reported. Note also, the predictive power (R ²) of the Actual and Intended models from Table 5 without the relevant bank account variable is 0.202 (0.238 with) and 0.231 (0.316 with), respectively.

A concern with the use of intention variables is whether they actually correspond with future action. Using the second wave of the PSED2 (undertaken a mean (median) of 12.5 (12) months after the first wave survey), we can observe the correspondence between intentions and actual actions over a future interval. Respondents were again asked if they intended to or had retained an accountant and lawyer. Of the 450 respondents whose first wave answers indicated they had not yet retained an accountant but who indicated whether they intended to retain one in the future, 71 (15.8%) had retained an accountant by the second wave. This proportion excludes the 150 ventures (33.3%) that still intended on retaining an accountant but had not done so by the second wave. For those that did (did not) intend to retain an accountant in the future (based on their first wave responses), accountant retention by the second wave was 20.1% (10.7%). This equates to an 88% greater likelihood of retaining an accountant by the second wave by those responding they would in the future (χ² = 7.44, p = 0.0057). Similarly, of the 487 respondents that had not yet retained a lawyer and provided information on subsequent lawyer use, 41 (8.4%) had retained a lawyer by the second wave, which excludes 112 ventures (30.0%) that still intended to retain a lawyer but had not done so by the second wave. Lawyer retention by the second wave was 14.3% (4.9%) for those that did (did not) intend to retain a lawyer in the future (χ² = 12.47, p = 0.0004). Thus, intended use is highly associated with future action, supporting the use of intention variables in this study.

We also examined the influence of number of owners using dummy variables representing: (a) one owner; (b) one owner and spouse; and (c) one owner and others. The results using these alternative variables resulted in the same inferences for the relationship between owners and accountant retention and are consequently not reported.

We also ran a series of accountant retention regressions including variables representing different business experience, namely: (a) work; (b) industry; (c) startup; (d) owner; and (e) manager. These tests provided little support for business experience affecting the study findings, with the explanatory power (R ²) of the models with the five experience variables increasing by only 0.02 for actual and 0.008 for intended retention. Running the regressions with just one experience variable at a time identified no significant associations between experience and accountant retention.

We also included an interaction between accounting experience and the presence of accounting role to investigate if the accounting experience of an owner dedicated to the accounting and finance function of the startup affected the result observed above. The coefficients for accounting role both with and without accounting experience were not significantly different from zero, or significantly different from each other.

To evaluate the robustness of the findings, we repeated the analyses using the population weights of all startups. The use of weights may be important, as these weights can address potential bias from non-coverage of those identified as nascent entrepreneurs in the screening survey but not selected for the phone survey, and non-response from those nascent entrepreneurs selected for the phone survey that did not complete the phone survey. In untabulated results, the economic magnitude and statistical significance of the explanatory variables are invariant to the use, or non-use, of population weights.