I think it’s a great idea to have a policy discussion based on real data! I’d like to make a few comments on the optimal design of corporate tax policy, informed by my PhD research at the London School of Economics. First, I think that reducing tax exemptions is generally a good idea. It will increase tax revenues, while spreading the tax burden more equally and reducing administrative costs, for both the firms and the tax authorities. Spreading the tax burden more equally will raise the perception that the system is fair and thus improve tax morale and compliance. Reducing exemptions may also reduce evasion. My research on UK corporations shows that small firms manipulate their deductions and costs to artificially decrease their taxable income and thus their tax liability. Some governments, such as those in Pakistan and Mexico, even limit the deductibility of costs in order to prevent firms from artificially under-reporting their tax revenue. Together a group of co-authors from LSE, we show that the government in Pakistan reduced tax evasion by imposing a turnover tax (rather than an income tax) on firms whose income tax liability falls below a certain threshold. As for tax evasion under the VAT, research by Dina Pomeranz (http://www.hbs.edu/faculty/Pages/item.aspx?num=43830) shows how audit threats can increase compliance, especially for transactions by retail companies and other firms that directly supply to the final consumer, as these transactions are not covered by the VAT paper trail. Audits (or threats thereof) also have important trickle down effects to suppliers further down the Vat chain. These multiplier effects should be taken account when choosing which firms to audit. In addition to the audit/enforcement policy, the design of the VAT system is a key determinant of compliance. Tanzania increased the VAT exemption threshold in 2004. This relieves many small firms from the burden of keeping records and filing VAT payments/claims. However, it also means that the VAT self-enforcement chain breaks down for value chains that include these small exempted firms. The tax authorities loose revenue not only from the newly exempted firms, but also from others that trade with them. Unfortunately, we do not have empirical evidence so far on the optimal VAT exemption threshold. By the way, I am not surprised that the tax revenue distribution in Tanzania is extremely skewed. Even in the UK, the largest 1% of firms account for 80% of corporate tax revenue (http://www.ifs.org.uk/publications/5885). This is due to the nature of the firm-size distribution, which is very skewed in developed economies and even more skewed in developing economies. Whether or not to attempt to increase tax revenue from the informal sector is of course an open debate… Personally, I think that taxing micro/informal sector firms should be attempted only if significant non-tax benefits are to be expected (i.e. firms are credit constraint and access to financial intermediation is conditional upon tax registration). Otherwise, the administrative costs of taxing small firms are likely to exceed the benefits in terms of increased tax revenue. As I have never lived in Tanzania, it seems inappropriate for me to comment on whether the government is raising enough revenue to fund a sufficient level of service provision. However, as was mentioned in a previous comment, the question is not only whether the revenue raised is enough to fund the desired level of service provision, but also how much of the revenue raised is actually channelled into service provision (rather than lost on the way in the pockets of corrupt bureaucrats or contractors). Closing those leaks (for instance through systematic publishing of records and accounts for all projects, and the involvement of civil society in social audits) may be easier and less distortionary to the economy than raising more tax revenue.