<p><em>Viv Davies interviews Robert Townsend for Vox</em></p>
<p><em>November 2010</em></p>
<p><em>Transcription of an VoxEU audio interview [http://www.voxeu.org/index.php?q=node/5937]</em></p>
<p><strong>Viv Davies</strong>: Hello, and welcome to Vox Talks, a series of audio interviews with leading economists from around the world. I'm Viv Davies from the Center for Economic Policy Research. It's the third of December, 2010. Early this week, Professor Robert Townsend of MIT was in London, presenting the Gorman Lectures at University College London. Stephen Yeo caught up with Professor Townsend immediately after one of the lectures, where he asked him to talk about his recently published book titled &quot;Households as Corporate Firms.&quot;</p>
<p><strong>Professor Robert Townsend</strong>: The book is basically trying to put households &ndash; often households running businesses &ndash; on the same footing as larger firms, in being able to use household data to create national income, and product accounts and flow of funds accounts. So, we had to think about the household as a firm. We already had the data. Unfortunately, we didn't come to a full realization of what we needed until after the questionnaires were already designed. So, in some respects, what we're trying to do is to provide a conceptual framework to organize the data into standard accounts. So, we would have, and now do have, the income statement, the statement of cash flow and the balance sheet for these households that we've been interviewing every month.</p>
<p><strong>Stephen Yeo</strong>: In the preface to the book, you gave as one of the arguments for doing things, this particular project, that it made it easier to define things properly, or to force yourself to define things properly. And, I suppose, by extension, to force everybody to define things in a broadly similar way. I can see that. I guess, to some extent, you're now wishing that you'd had this idea at the beginning of your... in 1997, rather than now, because I guess, to some extent, you're back fitting the data.</p>
<p><strong>Professor Townsen</strong>d: We can&rsquo;t re ask the questions that we've already asked all these years, but we certainly can identify the areas where better measurement would matter. Some of these distinctions have to do with, how do you measure income? It sounds like a simple question. If you're, on the other hand, thinking about the household as a firm, because, in fact, they have farm as well as non farm production activities, then one has to deal with something like gifts and transfers. Do we think of those as income?</p>
<p>Well, they do help you, in terms of your budget constraint. But they're not necessarily directly related to production activities. S now, when we see gifts in my data, we have to ask. Is this a gift, like an input into production? In which case, we have to value the gift.</p>
<p>They're really two simultaneous transactions. One is to get the gift and another is to use the income from the gift in production. So, that's a cost of production. If you ignored that gift, you would underestimate the cost.</p>
<p>And there are similar things on the output side. There are also some very hard things to estimate, like valuing the household's own labor in production. That does rank high on the list of things that matter. Unfortunately, it is something that's hard to estimate very accurately.</p>
<p><strong>Stephen</strong>: This is presumably more important the more the household is a productive entity in itself. Does it matter for a developed country household, where there are two adults going out to work and getting a week, monthly paycheck and that's it?</p>
<p><strong>Professor Townsend</strong>: No, it would matter much less for that. Yes, the point is that if one is thinking about and trying to model a developing country, where a substantial amount of economic activity is taking place within the household, the standard conceptualization of the national income accounts is actually to not do that. It's to make a separation between households and firms. Households are one actor or sector, firms are another. Households, like your laborer example just now, provide factors of production into firms and get compensated, and then turn around and buy goods that are produced by firms. But many households in developing countries are actually producing; some of the product they eat and some they sell on the market.</p>
<p><strong>Stephen</strong>: So, it matters to the extent that the household is actually a productive unit.</p>
<p><strong>Professor Townsend</strong>: Yes.</p>
<p><strong>Stephen</strong>: Rather than just supplying labor?</p>
<p><strong>Professor Townsend</strong>: That's right. So, this is a problem in almost all countries. In fact, ironically, in the creation of GDP and so on, national income, for India and all countries essentially, households are typically not very well measured. They're often even worse, the residual that makes things add up. So, savings - we argue the US isn't saving enough, just to name one issue. Those savings numbers typically come from the national income accounts, and so on. Whereas when you do household surveys, you get different answers.</p>
<p>This is a well known puzzle in development economics. And one of the points of having the accounts is to measure savings at the household level the way one ideally would have been measuring it in the national income and product accounts.</p>
<p><strong>Stephen</strong>: So, you're saying this is something that in principle you should do all the time. In practice, since national accounts were cooked up in the 1930s, 1940s, nobody cared very much about that issue, so they glossed over it. But it matters a lot in other countries.</p>
<p><strong>Professor Townsend</strong>: People were aware of it. Simon Kuznets and others talk about it. But they made kind of an odd decision, which is, if a household grows and eats its own food, well, that's not a market transaction. So, the national income accounts are supposed to be measuring market transactions, but the way they're used, they're measuring output. So it's a bit of a contradiction.</p>
<p><strong>Stephen</strong>: Right. And this is basically going back and saying this matters in certain countries. It matters all the time, but it matters a lot in certain countries.</p>
<p><strong>Professor Townsend</strong>: Thailand is a middle income country by now, and one would think that a large chunk of the national product is coming from large industries or even foreign owned Ford Motors or Nike producing in Thailand. But in fact, income from non farm, unincorporated enterprises is still the second largest category in the country. And it used to be the largest till the early `90s. Wage-earners is now the biggest single category. Something like the earnings or profits of incorporated enterprise are still only 20 percent of GDP. So, it's much smaller.</p>
<p>There's a tendency always, I think, to think about big as being significant. And you have a few, very big firms and you think they're producing a lot of the output, because after all, that's the modern sector and that must be where the action is. But then you have hundreds of thousands of small scale producers.</p>
<p>In fact, even as badly measured as the standard national income accounts are, even in those accounts, you end up with a big number for this small household sector.</p>
<p><strong>Stephen</strong>: Presumably, that's an even bigger issue in Africa, where that sector is even bigger.</p>
<p><strong>Professor Townsend</strong>: I would think that that is true, yes. So, hopefully, now having created the accounts and thought through some of the conceptual issues about what to do with home consumption and gifts and so on, we can provide some advice to other researchers. Hopefully, some people will take advantage of the lessons learned in the initial design of the surveys.</p>
<p><strong>Stephen</strong>: Very good, thank you very much.</p>
<p><strong>Professor Townsend</strong>: You're quite welcome. <br />
&nbsp;</p>
<p>&nbsp;</p>